Multi-index fusion calculation method for performance of water turbine in starting process of pumped storage unit
By employing methods such as data standardization and fuzzy membership functions, the problem of multi-index fusion calculation during the start-up of pumped storage turbine units was solved, enabling scientific and objective comprehensive performance evaluation, improving the scientific nature and adaptability of the evaluation, and supporting the optimization of start-up strategies and operational safety.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHINA YANGTZE POWER
- Filing Date
- 2026-02-12
- Publication Date
- 2026-06-09
AI Technical Summary
Existing technologies lack the integration and comprehensive evaluation of multi-source and multi-dimensional monitoring information systems during the start-up of pumped storage turbine units, resulting in biased and highly subjective evaluation results that affect the credibility and practicality of the evaluation. Furthermore, traditional methods are unable to scientifically quantify the mutual influence between various indicators and their contribution to the overall transient performance.
A multi-indicator fusion calculation method is constructed by employing data standardization, fuzzy membership function, information entropy and coefficient of variation calculation, combined with an adaptive weighting mechanism. The uncertainty of the indicators is handled by the fuzzy membership function, the weight of the indicators is objectively determined by the information entropy theory, and the adaptive mechanism is introduced to dynamically adjust the weight allocation, so as to achieve a scientific and objective comprehensive evaluation of the multi-indicator system.
It enables a scientific, objective, and comprehensive evaluation of the turbine startup process of pumped storage units, improves the scientific rigor and adaptability of the assessment, provides a systematic evaluation tool, and supports the refined optimization of startup strategies and operational safety.
Smart Images
Figure CN122173756A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of pumped storage power station technology, specifically to a method for calculating the multi-indicator performance of a pumped storage unit's turbine during startup. Background Technology
[0002] The startup transient process of a pumped-storage turbine, from standstill to rated speed, presents a critical challenge to the stability of the unit system due to its complex non-stationary dynamic characteristics. An inappropriate startup strategy can not only exacerbate fatigue damage to unit components but also induce hydraulic instability, abnormal shaft vibration, and other problems, threatening the long-term safe and stable operation of the unit.
[0003] Currently, performance evaluation of the startup process largely relies on single or a few indicators (such as total startup time, peak-to-peak vibration, etc.), lacking a systematic integration and comprehensive evaluation of multi-source, multi-dimensional monitoring information. Traditional methods struggle to objectively reflect the interrelationships between indicators and their combined contribution to overall transient performance, resulting in biased and highly subjective evaluation results that are difficult to support refined optimization of startup strategies. Furthermore, due to inconsistencies in dimensions, overlapping information, and ambiguity among indicators, direct weighting or comparison often lacks scientific basis, affecting the credibility and practicality of the evaluation.
[0004] Therefore, there is an urgent need for a startup performance evaluation method that can comprehensively process multi-source signals, adapt to non-stationary characteristics, and possess strong objectivity and adaptability, so as to scientifically quantify the comprehensive performance of the startup process and provide a reliable basis for optimizing the startup strategy and ensuring the safe operation of pumped storage units. Summary of the Invention
[0005] The technical problem to be solved by the present invention is to provide a multi-index fusion calculation method for the performance of pumped storage turbines during the start-up process, so as to realize the scientific and objective evaluation of the start-up performance of pumped storage turbines through a multi-index system.
[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A multi-indicator fusion calculation method for the turbine performance of pumped storage units during start-up includes the following steps: Step 1: Data Standardization; Since indicators typically have different dimensions and orders of magnitude, the raw data first needs to be standardized. For indicators where smaller is better, the following formula is used for standardization: ; in, x ij It is the original value of the j-th indicator of the i-th sample, max( x j ) and min(x j ) are the maximum and minimum values of the j-th indicator, respectively; after standardization y ij The value range is [0,1], and the larger the better; Step 2: Calculate the fuzzy membership degree; use the fuzzy membership degree function to transform the standardized data into a fuzzy evaluation matrix; Step 3: Calculate information entropy; information entropy is used to measure the amount of information in an indicator. Step 4: Calculate the coefficient of variation; the coefficient of variation is the ratio of the standard deviation to the mean, reflecting the relative degree of variation of the indicator. Step 5: Determine the adaptive weights; the adaptive weights are determined by the information entropy in Step 3 and the coefficient of variation in Step 4. Step 6: Calculate the overall performance score of the boot process; the overall performance score of the boot process is obtained by weighted summation, and a non-linear adjustment factor is introduced to enhance the discrimination.
[0007] In Step 2 above, the fuzzy membership function uses the triangular membership function, let... a j and b j Let be the threshold of the j-th indicator. Taking the quantiles of the standardized data, the membership function is defined as follows: .
[0008] The calculation process for information entropy in Step 3 above is as follows: For the j-th indicator, its information entropy is calculated as follows: First, calculate the feature weight of each sample under this indicator: ; Then, calculate the information entropy: ; Where n is the number of samples, when p ij When =0, it is stipulated that p ij ln p ij =0.
[0009] The calculation process for the coefficient of variation in Step 4 above is as follows: The coefficient of variation for the j-th indicator is: ; in, σ j It is the standard deviation of the j-th indicator. x ˉj It is the mean of the j-th indicator.
[0010] The calculation process for the adaptive weights in Step 5 above is as follows: First, calculate the weights based on information entropy: ; Then, calculate the weights based on the coefficient of variation: ; Finally, the two are combined to obtain the adaptive weights: ; Here, k represents the k-th indicator, and there are a total of m indicators.
[0011] The specific formula for the overall performance score of the above-mentioned boot process is as follows: ; in, β It is a non-linear adjustment parameter used to penalize indicators with low membership; the higher the score, the better the overall performance of the sample.
[0012] The raw data mentioned above includes multi-source signal indicators such as vibration, sway, pressure pulsation, and noise.
[0013] The above-mentioned pumped storage turbine start-up modes include open-loop start-up mode with large guide vane opening, open-loop start-up mode with small guide vane acceleration opening, closed-loop start-up mode with acceleration control, and open-loop + closed-loop start-up mode.
[0014] The above-mentioned open-loop start-up method with large guide vane opening is as follows: The first starting opening of the guide vane is the no-load limit, and the rate of change v of the guide vane control output is limited; when the unit speed exceeds the set ratio n1 of the rated speed, the guide vane opening is reduced back to the second starting opening; when the unit speed exceeds the set ratio n2 of the rated speed, n... 2＞ n1, the unit switches to no-load operation, and the guide vane opening is reduced back to the no-load opening.
[0015] The above-mentioned guide vane small acceleration opening open-loop start-up logic is as follows: the guide vane start-up opening is a preset proportion of the no-load opening, and it runs according to the set guide vane control output change rate. After a preset waiting time, it switches to the preset second start-up opening and maintains the corresponding change rate. When the unit speed reaches the preset proportion of the rated speed, it is pushed back to the preset no-load opening limit. When the unit speed reaches the target proportion of the rated speed, the speed governor engages frequency PID regulation. The open-loop start-up method with small acceleration of the guide vane includes several specific implementation forms. The difference between the different forms lies only in the specific values of the change rate of the guide vane control output and the waiting time. All of them follow the above core logic to achieve a smooth start-up of the unit.
[0016] The aforementioned acceleration control closed-loop start-up method is as follows: After receiving the start-up command, the speed governor sets the opening control to the preset start-up opening; from the start of unit rotation to the first preset proportional range of rated speed, the frequency setpoint tracks the unit frequency; when the unit speed reaches different preset proportional nodes of rated speed, the corresponding acceleration value is set respectively, and the acceleration control logic is operated according to the step-decreasing acceleration control logic until the frequency setpoint reaches the rated value; if the frequency setpoint reaches the preset threshold and the rated speed is not reached after maintaining the set time, it is directly set to the rated speed or grid frequency.
[0017] The specific logic of the above-mentioned open-loop + closed-loop start-up mode is as follows: After the speed controller receives the start-up command, it sets the opening degree control to the start-up opening degree and the frequency setpoint to the preset initial frequency value; the opening degree tracking opening degree control quickly opens to the start-up opening degree; when the frequency rises to the preset initial frequency value, the PID regulation is automatically engaged, and the frequency setpoint automatically increases from the preset initial frequency value to the rated frequency value according to the set ramp law. In the frequency tracking case, it is the grid frequency.
[0018] The above-mentioned data standardization application scenarios include the start-up process of pumped storage units under rated head conditions, where the rated head is a preset fixed value or dynamically determined based on actual power station parameters.
[0019] In the aforementioned step-decreasing acceleration control logic, the acceleration value at each stage is a positive value greater than 0, and the random group speed gradually decreases as it approaches the rated speed.
[0020] This invention discloses a multi-index fusion calculation method for the turbine performance of a pumped storage unit during start-up. It addresses the uncertainty of the indices through fuzzy membership functions, objectively determines the index weights using information entropy theory, and introduces an adaptive mechanism to dynamically adjust the weight allocation, thereby achieving a scientific and objective comprehensive evaluation of the multi-index system. By integrating multi-source signal indices such as vibration, sway, pressure pulsation, and noise, a systematic evaluation framework is constructed, providing an unprecedented scientific assessment tool for the unit's transient performance, exhibiting high objectivity and adaptability. Attached Figure Description
[0021] The present invention will be further described below with reference to the accompanying drawings and embodiments: Figure 1 This is a flowchart illustrating the invention method.
[0022] Figure 2 This is a schematic diagram showing the change of guide vane opening over time in the embodiment.
[0023] Figure 3 This is a schematic diagram illustrating the change of unit speed over time in the embodiment. Detailed Implementation
[0024] 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.
[0025] A multi-indicator fusion calculation method for the turbine performance of pumped storage units during start-up includes the following steps: Step 1: Data Standardization; Since indicators typically have different dimensions and orders of magnitude, the raw data first needs to be standardized. For indicators where smaller is better, the following formula is used for standardization:
[0026] in, x ij It is the original value of the j-th indicator of the i-th sample, max( x j ) and min( x j ) are the maximum and minimum values of the j-th indicator, respectively; after standardization y ij The value range is [0,1], and the larger the better; Step 2: Calculate the fuzzy membership degree; use the fuzzy membership degree function to transform the standardized data into a fuzzy evaluation matrix; Step 3: Calculate information entropy; information entropy is used to measure the amount of information in an indicator. Step 4: Calculate the coefficient of variation; the coefficient of variation is the ratio of the standard deviation to the mean, reflecting the relative degree of variation of the indicator. Step 5: Determine the adaptive weights; the adaptive weights are determined by the information entropy in Step 3 and the coefficient of variation in Step 4. Step 6: Calculate the overall performance score of the boot process; the overall performance score of the boot process is obtained by weighted summation, and a non-linear adjustment factor is introduced to enhance the discrimination.
[0027] In Step 2 above, the fuzzy membership function uses the triangular membership function, let... a j and b jLet be the threshold of the j-th indicator. Taking the quantiles of the standardized data, such as the 10th and 90th quantiles, the membership function is defined as follows: .
[0028] The calculation process for information entropy in Step 3 above is as follows: For the j-th indicator, its information entropy is calculated as follows: First, calculate the feature weight of each sample under this indicator: ; Then, calculate the information entropy: ; Where n is the number of samples, when p ij When =0, it is stipulated that p ij ln p ij =0.
[0029] The calculation process for the coefficient of variation in Step 4 above is as follows: The coefficient of variation for the j-th indicator is: ; in, σ j It is the standard deviation of the j-th indicator. x ˉ j It is the mean of the j-th indicator.
[0030] The calculation process for the adaptive weights in Step 5 above is as follows: First, calculate the weights based on information entropy: ; Then, calculate the weights based on the coefficient of variation: ; Finally, the two are combined to obtain the adaptive weights: ; Here, k represents the k-th indicator, and there are a total of m indicators.
[0031] The specific formula for the overall performance score of the above-mentioned boot process is as follows: ; in, β It is a non-linear adjustment parameter used to penalize indicators with low membership; the higher the score, the better the overall performance of the sample.
[0032] The raw data mentioned above includes multi-source signal indicators such as vibration, sway, pressure pulsation, and noise.
[0033] The above-mentioned pumped storage turbine start-up modes include open-loop start-up mode with large guide vane opening, open-loop start-up mode with small guide vane acceleration opening, closed-loop start-up mode with acceleration control, and open-loop + closed-loop start-up mode.
[0034] The above-mentioned open-loop start-up method with large guide vane opening is as follows: The first starting opening of the guide vane is the no-load limit, and the rate of change v of the guide vane control output is limited; when the unit speed exceeds the set ratio n1 of the rated speed, the guide vane opening is reduced back to the second starting opening; when the unit speed exceeds the set ratio n2 of the rated speed, n... 2＞ n1, the unit switches to no-load operation, and the guide vane opening is reduced back to the no-load opening.
[0035] The above-mentioned guide vane small acceleration opening open-loop start-up logic is as follows: the guide vane start-up opening is a preset proportion of the no-load opening, and it runs according to the set guide vane control output change rate. After a preset waiting time, it switches to the preset second start-up opening and maintains the corresponding change rate. When the unit speed reaches the preset proportion of the rated speed, it is pushed back to the preset no-load opening limit. When the unit speed reaches the target proportion of the rated speed, the speed governor engages frequency PID regulation. The open-loop start-up method with small acceleration of the guide vane includes several specific implementation forms. The difference between the different forms lies only in the specific values of the change rate of the guide vane control output and the waiting time. All of them follow the above core logic to achieve a smooth start-up of the unit.
[0036] The aforementioned acceleration control closed-loop start-up method is as follows: After receiving the start-up command, the speed governor sets the opening control to the preset start-up opening; from the start of unit rotation to the first preset proportional range of rated speed, the frequency setpoint tracks the unit frequency; when the unit speed reaches different preset proportional nodes of rated speed, the corresponding acceleration value is set respectively, and the acceleration control logic is operated according to the step-decreasing acceleration control logic until the frequency setpoint reaches the rated value; if the frequency setpoint reaches the preset threshold and the rated speed is not reached after maintaining the set time, it is directly set to the rated speed or grid frequency.
[0037] The specific logic of the above-mentioned open-loop + closed-loop start-up mode is as follows: After the speed controller receives the start-up command, it sets the opening degree control to the start-up opening degree and the frequency setpoint to the preset initial frequency value; the opening degree tracking opening degree control quickly opens to the start-up opening degree; when the frequency rises to the preset initial frequency value, the PID regulation is automatically engaged, and the frequency setpoint automatically increases from the preset initial frequency value to the rated frequency value according to the set ramp law. In the frequency tracking case, it is the grid frequency.
[0038] The above-mentioned data standardization application scenarios include the start-up process of pumped storage units under rated head conditions, where the rated head is a preset fixed value or dynamically determined based on actual power station parameters.
[0039] In the aforementioned step-decreasing acceleration control logic, the acceleration value at each stage is a positive value greater than 0, and the random group speed gradually decreases as it approaches the rated speed.
[0040] Example 1: This invention is illustrated using the specific start-up and shutdown scenarios of a pumped storage power station. This station undergoes thousands of start-ups and shutdowns annually and may operate under various suboptimal conditions. The reversible turbine's start-up and shutdown processes, as well as the transitions between operating conditions, are highly susceptible to unit vibration and sudden changes in dynamic stress, leading to serious hydraulic stability problems and equipment fatigue damage. Research on optimizing flexible start-up methods for pumped storage power station units has significant guiding value for the industry, providing a strong reference for the development of subsequent start-up methods for power stations, further improving the safety and stability of pumped storage power station equipment and facilities, and yielding significant economic and social benefits.
[0041] The power station currently uses a start-up method with a large guide vane opening for the turbines, referred to as Method 1: the first start-up opening of the guide vanes is at the no-load limit, and the no-load opening... 120%+3%, the rate of change v of the guide vane control output is limited to about 5% / s. When the unit speed is greater than 95% of the rated speed, the pressure returns to the second starting opening, no-load opening. 110%+2%, when the unit speed is greater than 96% of the rated speed, the unit switches to no-load operation and the guide vane opening is reduced back to the no-load opening.
[0042] To provide a more reasonable boot-up pattern, this invention also includes: 1. Open-loop start with small guide vane acceleration opening, i.e., method 2.1: The guide vane starting opening is approximately 60% of the no-load opening, and the rate of change of the guide vane control output... v Limit the opening speed to approximately 0.7% / s, wait about 5 seconds, and then switch to the no-load opening. Y The second starting opening is 0 plus 8%, and the rate of change of the guide vane control output is... v Limited to approximately 0.7% / s, when the unit speed reaches about 90% of the rated speed, the pressure returns to the no-load opening limit, which is 5% larger than the no-load opening. The rate of change of the guide vane control output is... v Limited to approximately 0.7% / s, when the unit speed reaches about 95% of the rated speed, the speed governor engages frequency PID regulation.
[0043] 2. Open-loop start with small guide vane acceleration opening, i.e., method 2.4: The guide vane starting opening is approximately 60% of the no-load opening, and the rate of change of the guide vane control output... v Limit the opening speed to approximately 1.6% / s, wait about 7 seconds, and then switch to the no-load opening. Y The second starting opening is 0 plus 8%, and the rate of change of the guide vane control output is... v Limited to approximately 1.6% / s, when the unit speed reaches about 90% of the rated speed, the pressure returns to the no-load opening limit, which is 5% larger than the no-load opening. The rate of change of the guide vane control output... v Limited to around 1.6% / s, when the unit speed reaches about 95% of the rated speed, the speed governor engages frequency PID regulation.
[0044] 3. Open-loop start with small guide vane acceleration opening, i.e., mode 2.5: The guide vane starting opening is approximately 60% of the no-load opening, and the rate of change of the guide vane control output... v Limit the opening speed to approximately 1.6% / s, wait about 5 seconds, and then switch to the no-load opening. Y The second starting opening is 0 plus 8%, and the rate of change of the guide vane control output is... v Limited to approximately 1% / s, when the unit speed reaches about 90% of the rated speed, the pressure returns to the no-load opening limit, which is 5% larger than the no-load opening. The rate of change of the guide vane control output is... v Limited to approximately 1% / s, when the unit speed reaches about 95% of the rated speed, the speed governor engages frequency PID regulation.
[0045] 4. Acceleration control closed-loop start, i.e., method 3: After receiving the start-up command, the speed governor sets the opening control to the starting opening and the no-load opening +3%. From the start of the unit's rotation to 20% of the rated speed, the frequency setpoint tracks the unit's frequency. Before the unit speed reaches approximately 85% of the rated speed, the unit acceleration is set to 2Hz / s; before the unit speed reaches approximately 95% of the rated speed, the unit acceleration is set to 1Hz / s; after the unit speed reaches 95% of the rated speed, the unit acceleration is set to 0.5Hz / s, until the frequency setpoint reaches the rated value. If the frequency setpoint is already greater than 49.7 Hz, it is maintained for more than 5 seconds. If it still does not reach the rated speed of 50Hz, the rated speed of 50Hz or the grid frequency is directly applied.
[0046] 5. Open-loop + closed-loop start-up, i.e., method 4: After receiving the start-up command, the speed controller sets the opening control to the start-up opening, and the frequency setpoint is set to 40Hz. The opening tracking control quickly opens to the start-up opening. When the frequency rises to 40Hz, the PID regulation is automatically engaged, and the frequency setpoint starts to automatically increase from 40Hz to 50Hz according to the set ramp. In the frequency tracking mode, the grid frequency is used, and the unit frequency will enter 50Hz relatively smoothly.
[0047] During the unit startup process, the entire process of changes in unit vibration, sway, pressure pulsation, and noise is simultaneously collected. Through testing, the unit's vibration, sway, pressure pulsation, and noise levels under this startup mode are understood.
[0048] Field tests were conducted on the hydropower units of the power station to study the stability of the units under the above six start-up methods. Field tests were also conducted on a pumped storage unit under the rated head for the start-up method.
[0049] II. Changes in guide vane opening and rotational speed A pumped-storage unit underwent a rated head start-up test. The test included (1) open-loop start-up with large guide vane opening; and (2) open-loop start-up with small guide vane opening. There were three types of start-up, named Mode 2.1, Mode 2.4, and Mode 2.5 respectively. The rate of change of the guide vane control output was measured in each of the three modes. v The changes in the guide vane opening and rotational speed of the unit are as follows: (3) Acceleration control closed-loop start-up mode; (4) Open-loop + closed-loop start-up mode, etc., when there are six start-up modes, the changes are as follows: Figures 2-3 As shown.
[0050] Figure 2 The figure shows the change in guide vane opening over time under six start-up modes. As can be seen from the figure, the final opening reaches a stable no-load opening under the six modes, with the total time taken, from smallest to largest, being: Mode 2.5, Mode 2.4, Mode 1, Mode 3, Mode 4, and Mode 2.1. Modes 1 and 4 show the largest changes in guide vane opening in the first stage. In Mode 1, the guide vane opening remains in the first position for approximately 30 seconds before opening to other openings. In Mode 2.1, the guide vane opening remains in the first position for approximately 5 seconds before opening to other openings. In Mode 2.4, the guide vane opening remains in the first position for approximately 7 seconds before opening to other openings. In Mode 2.5, the guide vane opening remains in the first position for approximately 5 seconds before opening to other openings. In Mode 3, the guide vane opening remains in the first position for approximately 10 seconds before opening to other openings. In Mode 4, the guide vane opening remains in the first position for approximately 23 seconds before opening to other openings. Method 1 results in the longest dwell time at approximately 17.5% of the guide vane opening.
[0051] Figure 3The figure shows the changes in unit speed over time under six different start-up tests. As can be seen from the figure, the time required for the unit to reach 95% speed from rest, from shortest to longest, is as follows: Mode 2.5, Mode 2.4, Mode 1, Mode 3, Mode 4, and Mode 2.1. Mode 2.5 shows the fastest speed increase. Mode 2.1 shows a slow speed increase in the first 17 seconds, Mode 2.4 shows a slow speed increase in the first 13 seconds, and Mode 2.5 shows a slow speed increase in the first 10 seconds. After the guide vane opening is increased for the second time, the speed increases rapidly. Mode 4 shows a relatively gradual speed change.
[0052] This invention has the following innovative points: 1. Systematic multi-index fusion: It creatively integrates multi-source and multi-dimensional monitoring signals such as vibration, sway, pressure pulsation, and noise to construct a complete comprehensive evaluation index system; 2. Introduction of fuzzy mathematics theory: By introducing fuzzy membership functions, such as triangular membership functions, the standardized precise data is transformed into a fuzzy evaluation matrix, enhancing the rationality and fault tolerance of the evaluation results; 3. Objective adaptive weighting method based on information entropy and coefficient of variation: Combining information entropy and coefficient of variation, the constructed adaptive weights can dynamically and objectively reflect the true importance of each index in the comprehensive evaluation, significantly improving the scientificity and objectivity of the evaluation; 4. Construction of a nonlinear comprehensive scoring model: By introducing a nonlinear adjustment factor, the differences in comprehensive scores among different schemes are amplified. This makes the evaluation results more discriminative and sensitive, facilitating the clear identification of the optimal solution among multiple alternative schemes.
Claims
1. A method for calculating multiple performance indicators of a pumped storage turbine during start-up, characterized in that... Includes the following steps: Step 1: Data Standardization; Since indicators typically have different dimensions and orders of magnitude, the raw data first needs to be standardized. For indicators where smaller is better, the following formula is used for standardization: ； in, x ij It is the original value of the j-th indicator of the i-th sample, max( x j ) and min( x j ) are the maximum and minimum values of the j-th indicator, respectively; after standardization y ij The value range is [0,1], and the larger the better; Step 2: Calculate the fuzzy membership degree; use the fuzzy membership degree function to transform the standardized data into a fuzzy evaluation matrix; Step 3: Calculate information entropy; information entropy is used to measure the amount of information in an indicator. Step 4: Calculate the coefficient of variation; the coefficient of variation is the ratio of the standard deviation to the mean, reflecting the relative degree of variation of the indicator. Step 5: Determine the adaptive weights; the adaptive weights are determined by the information entropy in Step 3 and the coefficient of variation in Step 4. Step 6: Calculate the overall performance score of the boot process; the overall performance score of the boot process is obtained by weighted summation, and a non-linear adjustment factor is introduced to enhance the discrimination.
2. The method for multi-index fusion calculation of the turbine performance during the start-up process of a pumped storage unit as described in claim 1, is characterized in that... In Step 2, the fuzzy membership function adopts the triangular membership function, let... a j and b j Let be the threshold of the j-th indicator. Taking the quantiles of the standardized data, the membership function is defined as follows: 。 3. The method for multi-index fusion calculation of the turbine performance during the start-up process of a pumped storage unit as described in claim 1, characterized in that, The calculation process of information entropy in Step 3 is as follows: For the j-th indicator, its information entropy is calculated as follows: First, calculate the feature weight of each sample under this indicator: ； Then, calculate the information entropy: ； Where n is the number of samples, when p ij When =0, it is stipulated that p ij ln p ij =0.
4. The method for multi-index fusion calculation of the turbine performance during the start-up process of a pumped storage unit as described in claim 3, is characterized in that... The calculation process for the coefficient of variation in Step 4 is as follows: The coefficient of variation for the j-th indicator is: ； in, σ j It is the standard deviation of the j-th indicator. x ˉ j It is the mean of the j-th indicator.
5. The method for multi-index fusion calculation of the turbine performance during the start-up process of a pumped storage unit as described in claim 4, characterized in that, The calculation process for the adaptive weights in Step 5 is as follows: First, calculate the weights based on information entropy: ； Then, calculate the weights based on the coefficient of variation: ； Finally, the two are combined to obtain the adaptive weights: ； Here, k represents the k-th indicator, and there are a total of m indicators.
6. The method for multi-index fusion calculation of the turbine performance during the start-up process of a pumped storage unit as described in claim 5, is characterized in that... The specific formula for the comprehensive performance score of the boot process is as follows: ； in, β It is a non-linear adjustment parameter used to penalize indicators with low membership; the higher the score, the better the overall performance of the sample.
7. The method for multi-index fusion calculation of the turbine performance during the start-up process of a pumped storage unit as described in claim 1, characterized in that, The raw data includes multi-source signal indicators such as vibration, sway, pressure pulsation, and noise.
8. The method for multi-index fusion calculation of the turbine performance during the start-up process of a pumped storage unit as described in claim 1, characterized in that, The pumped storage unit turbine start-up modes include open-loop start-up mode with large guide vane opening, open-loop start-up mode with small guide vane acceleration opening, closed-loop start-up mode with acceleration control, and open-loop + closed-loop start-up mode.
9. The method for multi-index fusion calculation of the turbine performance during the start-up process of a pumped storage unit as described in claim 8, characterized in that, The open-loop start-up method with large start-up opening of the guide vane is specifically as follows: The first starting opening of the guide vane is the no-load limit, and the rate of change v of the guide vane control output is limited; when the unit speed exceeds the set ratio n1 of the rated speed, the guide vane opening is reduced back to the second starting opening; when the unit speed exceeds the set ratio n2 of the rated speed, n... 2＞ n1, the unit switches to no-load operation, and the guide vane opening is reduced back to the no-load opening.
10. The method for multi-index fusion calculation of the turbine performance during the start-up process of a pumped storage unit as described in claim 8, characterized in that, The logic of the open-loop start-up mode with small acceleration opening of the guide vane is as follows: the starting opening of the guide vane is a preset proportion of the no-load opening, and it runs according to the set guide vane control output change rate. After a preset waiting time, it switches to the preset second starting opening and maintains the corresponding change rate. When the unit speed reaches the preset proportion of the rated speed, it is pushed back to the preset no-load opening limit. When the unit speed reaches the target proportion of the rated speed, the speed governor engages frequency PID regulation. The open-loop start-up method with small guide vane opening includes various specific implementation forms. The difference between the different forms lies only in the specific values of the change rate of the guide vane control output and the waiting time. All of them follow the above core logic to achieve smooth unit start-up.
11. The method for multi-index fusion calculation of the turbine performance during the start-up process of a pumped storage unit as described in claim 8, characterized in that, The specific acceleration control closed-loop start-up method is as follows: After receiving the start-up command, the speed governor sets the opening control to a preset start-up opening; from the start of unit rotation to the first preset proportional range of rated speed, the frequency setpoint tracks the unit frequency; when the unit speed reaches different preset proportional nodes of rated speed, the corresponding acceleration value is set respectively, and the acceleration control logic is operated according to a step-decreasing manner until the frequency setpoint reaches the rated value; if the frequency setpoint reaches the preset threshold and the rated speed is not reached after maintaining the set time, it is directly set to the rated speed or grid frequency.
12. The method for multi-index fusion calculation of the turbine performance during the start-up process of a pumped storage unit as described in claim 8, characterized in that, The specific logic of the open-loop + closed-loop start-up mode is as follows: After the speed controller receives the start-up command, it sets the opening degree control to the start-up opening degree and the frequency setpoint to the preset initial frequency value; the opening degree tracking opening degree control quickly increases to the start-up opening degree; when the frequency rises to the preset initial frequency value, the PID regulation is automatically engaged, and the frequency setpoint automatically increases from the preset initial frequency value to the rated frequency value according to the set ramp law. In the frequency tracking case, it is the grid frequency.
13. The method for multi-index fusion calculation of the turbine performance during the start-up process of a pumped storage unit as described in claim 1, characterized in that, The application scenarios for the data standardization mentioned include the start-up process of pumped storage units under rated head conditions, where the rated head is a preset fixed value or dynamically determined based on actual power station parameters.
14. The method for multi-index fusion calculation of the turbine performance during the start-up process of a pumped storage unit as described in claim 11, characterized in that, In the aforementioned step-decreasing acceleration control logic, the acceleration value at each stage is a positive value greater than 0, and the random group speed gradually decreases as it approaches the rated speed.