A method for judging mechanical fault of hydroelectric generating unit based on speed fluctuation characteristics

Abstract

The present application relates to the technical field of hydroelectric unit fault determination, and particularly relates to a hydroelectric unit mechanical fault determination method based on speed fluctuation characteristics, comprising: acquiring unit operation state data in real time, identifying an adjusting action end time, intercepting a speed time sequence data segment between the adjusting action end time and a time when the unit enters a complete steady state from a speed signal to obtain a mechanical inertia response segment; performing reference steady state correction and disturbance elimination processing to obtain a corrected mechanical inertia response segment; extracting a speed fluctuation characteristic set; comparing the speed fluctuation characteristic set with a normal speed fluctuation reference characteristic to generate a mechanical fluctuation deviation result, identifying a mechanical abnormality type according to the mechanical fluctuation deviation result, and combining the mechanical abnormality type to generate a hydroelectric unit mechanical fault determination result. The present application can reflect potential mechanical faults such as rotational inertia abnormality, damping abnormality and local mechanical jamming, and improve the early identification ability and dynamic diagnosis ability of hydroelectric unit mechanical faults.

Description

Technical Field

[0001] This invention relates to the field of hydropower unit fault diagnosis technology, and in particular to a method for diagnosing mechanical faults in hydropower units based on speed fluctuation characteristics. Background Technology

[0002] With the long-term high-load operation and frequent participation in power grid frequency regulation and peak shaving of hydropower units, rotating parts, shaft structures, and speed-regulating mechanical mechanisms are prone to mechanical abnormalities such as changes in rotational inertia, damping degradation, misalignment, imbalance, and local friction jamming. These mechanical abnormalities typically do not manifest as obvious shutdown faults in their early stages, but they gradually affect the unit's operational stability, leading to increased vibration, decreased mechanical efficiency, and shortened service life. Therefore, timely and accurate fault diagnosis of the hydropower unit's mechanical condition is necessary. Existing methods for detecting mechanical faults in hydropower units mostly rely on vibration sensors, temperature sensors, or steady-state operating parameters for analysis, primarily monitoring steady-state characteristics after obvious mechanical abnormalities have already formed. However, during the mechanical inertia recovery process after the unit's adjustment actions are completed, the speed signal actually includes a large amount of dynamic fluctuation information that reflects changes in mechanical state, such as changes in speed stabilization rate, changes in fluctuation attenuation capability, and local abnormal oscillations. Existing technologies typically lack specialized analysis mechanisms for this type of mechanical inertial response process, making it difficult to identify some early mechanical anomalies in a timely manner. In addition, since the speed fluctuation characteristics under different loads, heads, and grid connection conditions are themselves quite different, existing technologies using fixed thresholds or single fluctuation indicators for fault judgment are easily affected by changes in operating conditions, external power grid disturbances, and instantaneous noise interference, leading to misjudgments or missed judgments. Existing methods lack the ability to effectively distinguish the dynamic fluctuation differences between different types of mechanical anomalies, making it difficult to further achieve mechanical anomaly type identification and fault severity classification. Summary of the Invention

[0003] This invention provides a method for determining mechanical faults in hydropower units based on speed fluctuation characteristics, which can identify and classify mechanical faults in hydropower units based on dynamic characteristics of speed fluctuations.

[0004] A method for determining mechanical faults in hydropower units based on rotational speed fluctuation characteristics includes the following steps: S1: During the operation of the hydropower unit, the unit's operating status data is acquired in real time, the end time of the adjustment action is identified, and the speed time sequence data segment between the end time of the adjustment action and the moment when the unit enters a completely steady state is extracted, using the end time of the adjustment action as the starting point, to obtain the mechanical inertial response segment. S2: Perform benchmark steady-state correction and disturbance removal processing on the mechanical inertial response segment to obtain the corrected mechanical inertial response segment; extract the rotational speed fluctuation feature set based on the corrected mechanical inertial response segment; S3: Compare the speed fluctuation feature set with the normal speed fluctuation benchmark feature under the corresponding operating conditions of the hydropower unit to generate mechanical fluctuation deviation results, identify the mechanical anomaly type based on the mechanical fluctuation deviation results, and generate the mechanical fault judgment result of the hydropower unit in combination with the mechanical anomaly type.

[0005] Optionally, S1 specifically includes: S11: Synchronously acquire unit operating status data at a sampling frequency capable of completely capturing transient changes in rotational speed, and store the acquired signals in time alignment; the unit operating status data includes rotational speed signal, guide vane opening command signal, grid connection frequency signal, and load status signal; S12: Real-time monitoring of the rate of change of the guide vane opening command signal, the grid connection frequency signal, and the load status signal to determine the end time of the adjustment action; S13: Taking the end time of the adjustment action as the starting point of the interception, the sliding window fluctuation amplitude of the speed signal relative to the estimated steady-state speed is calculated in real time. When the sliding window fluctuation amplitude is less than the preset speed steady-state determination threshold in multiple consecutive sliding windows, the unit is determined to have entered the complete steady-state moment. The speed time sequence data segment from the interception starting point to the complete steady-state moment is extracted as the mechanical inertial response segment.

[0006] Optionally, the speed signal is used to characterize the real-time speed change of the rotating components of the unit, the guide vane opening command signal is used to characterize the change of the control command of the speed regulation system on the guide vane opening, the grid connection frequency signal is used to characterize the frequency stability of the power grid currently connected to the unit, and the load status signal is used to characterize whether the unit load has changed or been disturbed.

[0007] Optionally, the conditions for determining the end time of the adjustment action specifically include: Condition 1: The rate of change is consistently less than the pre-calibrated governor command stability threshold and the guide vane opening command signal does not exceed the adjustment dead zone change in subsequent consecutive sampling periods; Condition 2: The deviation between the grid connection frequency signal and the rated frequency is continuously less than the preset frequency tolerance threshold, and the load status signal remains unchanged for multiple consecutive sampling periods; If both conditions one and two are met, then the current time will be determined as the end time of the adjustment action.

[0008] Optionally, S2 specifically includes: S21: Perform a reference steady-state correction on the rotational speed time series data in the mechanical inertial response segment. The reference steady-state correction is to fit the overall rotational speed change trend in the mechanical inertial response segment and subtract the overall change trend from the original rotational speed signal to obtain a correction signal after eliminating the reference offset. Perform disturbance removal processing on the correction signal to obtain the corrected mechanical inertial response segment. S22: Extract a speed fluctuation feature set based on the corrected mechanical inertial response segment. The speed fluctuation feature set includes speed stabilization and persistence features, fluctuation attenuation features, residual fluctuation features near steady state, positive and negative fluctuation asymmetry features, and local reverse fluctuation features.

[0009] Optionally, the disturbance removal process includes identifying and removing abnormal amplitude points caused by instantaneous sensor interference or external random impact, and interpolating and repairing the removed points to obtain the corrected mechanical inertial response segment.

[0010] Optionally, the sustained characteristic of speed stabilization includes the length of time from the end of the adjustment action until the speed fluctuation amplitude first enters and remains within the steady-state allowable band; the fluctuation attenuation characteristic includes the attenuation rate of the peak and trough values ​​of speed fluctuation over time or the slope of the envelope in the corrected mechanical inertial response segment; the residual fluctuation characteristic near the steady state includes the high-frequency residual fluctuation amplitude or energy of the speed relative to the steady-state mean over a period of time after the unit enters a fully steady state; the asymmetric positive and negative fluctuation characteristic includes the degree of difference between the cumulative amount of positive and negative speed fluctuation amplitudes in the mechanical inertial response segment; the local reverse fluctuation characteristic includes a local speed recovery or oscillation amplitude that occurs opposite to the main attenuation direction under the overall attenuation trend.

[0011] Optionally, S3 specifically includes: S31: Obtain a normal speed fluctuation benchmark feature set that matches the current operating conditions of the hydropower unit. Compare each feature in the speed fluctuation feature set with the corresponding normal speed fluctuation benchmark feature one by one, calculate the difference between each feature, and when the difference of any feature exceeds the corresponding preset deviation tolerance range, mark the corresponding feature as an abnormal fluctuation feature. Collect all abnormal fluctuation features and differences to generate the mechanical fluctuation deviation result. S32: Identify the mechanical anomaly type based on the mechanical fluctuation deviation result, and map it to the pre-established anomaly type discrimination rule base according to the combination pattern of different abnormal fluctuation characteristics and the relative magnitude of the difference. S33: Combining the mechanical anomaly type and the characteristic difference level in the mechanical fluctuation deviation result, a comprehensive mechanical fault judgment result for the hydropower unit is generated. The mechanical fault judgment result includes the fault type, fault severity level, and corresponding maintenance suggestion label.

[0012] Optionally, the normal speed fluctuation reference feature set includes at least the normal speed stabilization and persistence feature, the normal fluctuation attenuation feature, the residual fluctuation feature near the normal steady state feature, the normal positive and negative fluctuation asymmetry feature, and the normal local reverse fluctuation feature.

[0013] Optionally, the anomaly type discrimination rule base includes the correspondence between feature combinations and mechanical anomaly types, including rotational inertia anomaly type, damping anomaly type, misalignment or imbalance type, and local friction or jamming type.

[0014] The beneficial effects of this invention are: This invention automatically extracts mechanical inertial response segments after the adjustment action ends and constructs a mechanical fluctuation analysis mechanism based on the speed recovery process. It can extract mechanical state information from the transient inertial recovery process, which is difficult to reflect by traditional steady-state monitoring. Compared with detection methods that rely solely on steady-state vibration or temperature parameters, it can reflect potential mechanical faults such as abnormal rotational inertia, abnormal damping, and local mechanical jamming, thereby improving the early identification and dynamic diagnosis capabilities of hydropower unit mechanical faults.

[0015] This invention constructs a speed fluctuation feature set by performing benchmark steady-state correction, abnormal amplitude point elimination, and multi-dimensional speed fluctuation feature extraction on mechanical inertial response segments. This set includes speed stabilization persistence features, fluctuation attenuation features, residual fluctuation features near steady state, positive and negative fluctuation asymmetry features, and local reverse fluctuation features. It can perform fine-grained characterization of the dynamic differences of different types of mechanical anomalies in the inertial recovery process. It can effectively distinguish different mechanical anomaly states such as insufficient inertia, damping degradation, misalignment, imbalance, and local friction, thereby improving the accuracy and diagnostic precision of mechanical fault type identification.

[0016] This invention establishes a normal speed fluctuation benchmark feature set that matches the operating conditions, and combines it with an anomaly type discrimination rule base to comprehensively analyze the combination patterns and feature differences of abnormal fluctuation features. This achieves integrated processing for mechanical anomaly type identification, fault severity classification, and maintenance suggestion generation. It not only avoids interference from differences in normal operating conditions under different loads, heads, and grid connection states on fault judgment results, but also dynamically outputs maintenance suggestions such as continued operation observation, planned maintenance, or shutdown maintenance based on the degree of anomaly. This improves the engineering applicability of hydropower unit mechanical fault judgment results and the operation and maintenance decision support capability. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only for this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the method flow according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the mechanical determination in an embodiment of the present invention. Detailed Implementation

[0019] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. For some well-known technologies, those skilled in the art may also use other alternative methods to implement the invention. Moreover, the accompanying drawings are only for more specific description of the embodiments and are not intended to specifically limit the present invention.

[0020] like Figure 1 - Figure 2 As shown, a method for determining mechanical faults in hydropower units based on speed fluctuation characteristics includes the following steps: S1: During the operation of the hydropower unit, the unit's operating status data is acquired in real time, the end time of the adjustment action is identified, and the speed time sequence data segment between the end time of the adjustment action and the moment when the unit enters a completely steady state is extracted, thus obtaining the mechanical inertial response segment. S1 specifically includes: S11. During the operation of the hydropower unit, the unit monitoring system synchronously collects speed signals, guide vane opening command signals, grid connection frequency signals, and load status signals using a unified clock reference. Among them, the speed signal is used to characterize the real-time speed change of the unit's rotating components; the guide vane opening command signal is used to characterize the change of the speed control system's control command for the guide vane opening; the grid connection frequency signal is used to characterize the frequency stability of the power grid currently connected to the unit; and the load status signal is used to characterize whether the unit's load has switched or experienced disturbances. The sampling frequency is set according to the transient change characteristics of the unit's speed, ensuring that the sampling frequency meets the requirement of fully capturing the details of speed fluctuations. After the acquisition of each signal is completed, it is time-aligned and stored according to a unified timestamp, forming a multi-source synchronous operating status data sequence. Wherein, the sampling frequency satisfies ; in, Indicates the signal sampling frequency. This represents the highest target frequency component in the unit speed fluctuation.

[0021] S12: Calculate the rate of change of the guide vane opening command signal in real time, obtain the rate of change of the guide vane opening command between adjacent sampling times, and identify whether the speed regulation action has ended based on the continuous and stable change rate. The rate of change of guide vane opening command is expressed as: ; in, Indicates the first The rate of change of guide vane opening command at each sampling time. Indicates the first The guide vane opening command value at each sampling time. Indicates the sampling time interval.

[0022] When the guide vane opening command change rate continuously meets Furthermore, the guide vane opening command signal satisfies the following conditions in multiple subsequent consecutive sampling periods. ; At the same time, the grid connection frequency signal meets ; Furthermore, if the load status signal remains without load switching indicator for multiple consecutive sampling periods, it is determined that the current speed regulation process has ended, and the current time is determined as the end time of the regulation action.

[0023] in, The stable threshold of the speed governor command is set to 0.05% to 0.2% of the rated guide vane opening, preferably 0.1%. When the rate of change of the guide vane opening command is lower than this range, the speed governor output has basically entered the stable adjustment stage. Further changes mainly come from controller fine-tuning or signal noise, and no longer represent effective adjustment actions. This indicates the threshold for dead zone variation, set at 0.1% to 0.5% of the rated guide vane opening, with a value of 0.2%. Hydropower unit speed control systems typically exhibit mechanical hysteresis, hydraulic micro-oscillations, and control dead zones near steady state. Guide vane opening fluctuations smaller than this range do not affect the unit's mechanical inertial response characteristics and can be considered normal, minor fluctuations under steady-state conditions. Indicates the first The grid connection frequency corresponding to each sampling time is the real-time operating frequency of the power grid, with values ​​ranging from 49.8Hz to 50.2Hz. This indicates the rated frequency, with a value of 50. 50 is the standard operating frequency of the power grid. When the real-time frequency deviates from the rated frequency by a small amount, it indicates that the power grid is operating stably. The change in speed mainly reflects the mechanical inertial response of the unit itself, rather than external power grid disturbances. The frequency tolerance threshold is defined as 0.02 to 0.10 times the rated frequency, preferably 0.05. When the hydropower unit is connected to a 50 kWh power grid, the allowable grid-connected frequency fluctuation range is preferably 49.95 to 50.05 kWh. This is because when the grid frequency deviation is within this range, the overall operation of the power grid is basically stable, and speed fluctuations are mainly caused by the unit's own mechanical inertial response, the speed regulation system's stabilization process, and the mechanical structure status. This avoids interference from external grid frequency disturbances on the extraction of mechanical fluctuation characteristics. If the frequency deviation exceeds this range, grid-side disturbance components will be significantly superimposed on the speed changes, easily reducing the accuracy of mechanical fault determination results. Indicates the sequence number of the continuous sampling period.

[0024] S13: Using the end of the adjustment action as the starting point for extracting the mechanical inertial response segment, a sliding window fluctuation analysis is performed on the subsequent speed signal; firstly, the estimated steady-state speed is estimated based on historical stable operating data corresponding to the current operating condition, and the estimated steady-state speed is used as the reference benchmark for speed fluctuation; subsequently, within a length of... The speed fluctuation amplitude is calculated within the sliding window and expressed as: ; in, Indicates the first The speed fluctuation range within each sliding window Indicates the first The data set corresponding to each sliding window Indicates the first [number]th ... The rotational speed value corresponding to each sampling point.

[0025] When multiple consecutive sliding windows satisfy ; Then it is determined that the current unit has entered a fully steady state. in, The steady-state speed determination threshold is 0.02% to 0.10% of the rated speed, with a value of 0.05%. For example, for a hydropower unit with a rated speed of 300 rpm, the steady-state speed determination threshold can be set to 0.06 to 0.30, preferably 0.15. When the speed fluctuation within the sliding window is below this range, the unit speed has basically completed the inertial recovery process. The remaining fluctuations mainly come from measurement noise, minor hydraulic disturbances, and fine-tuning behavior of the speed control system, and no longer show obvious mechanical transient response characteristics. If the threshold is set too small, it is easy to cause a delay in steady-state determination due to normal micro-oscillations. If the threshold is set too large, it may prematurely truncate the mechanical inertial response segment before the unit is fully stable, affecting the accuracy of subsequent mechanical fluctuation feature extraction.

[0026] After determining the moment of complete steady state, all rotational speed time series data from the end of the adjustment action to the moment of complete steady state are extracted to form the corresponding mechanical inertial response segment, which is used for subsequent extraction of mechanical fluctuation features and mechanical fault judgment and processing.

[0027] S2: Perform baseline steady-state correction and disturbance removal on the mechanical inertial response segment to obtain the corrected mechanical inertial response segment; extract the speed fluctuation feature set based on the corrected mechanical inertial response segment; S2 specifically includes: S21: Perform benchmark steady-state correction processing on the speed time series data in the mechanical inertial response segment to eliminate the overall speed drift effect of the unit under different operating conditions; S211: Fit the overall trend of the original rotational speed sequence in the mechanical inertial response segment to obtain the corresponding trend reference curve; that is, select a fixed-length time window near each sampling point, calculate the average value of all rotational speed values ​​within the window, and use the average value as the trend value corresponding to the current moment; as the time window slides continuously along the rotational speed sequence, a continuous trend reference curve can be formed, which characterizes the low-frequency overall trend of rotational speed during the inertial recovery process; subsequently, subtract the overall trend from the original rotational speed signal to obtain the corrected signal after eliminating the reference offset; The reference steady-state correction is expressed as: ; in, Indicates the first The correction signal corresponding to each sampling time. This represents the original rotational speed signal. This represents the overall trend value obtained from the fitting.

[0028] S212: After completing the reference steady-state correction, the correction signal is subjected to disturbance removal processing. Abnormal amplitude points in the correction signal caused by sensor instantaneous interference, data acquisition jitter or external random impact are detected. When the deviation between a certain sampling point and its neighborhood mean exceeds the abnormal judgment threshold, the corresponding sampling point is identified as an abnormal amplitude point. Abnormal amplitude points are identified as follows: ; After identifying abnormal amplitude points, the abnormal points are removed, and interpolation is performed using adjacent valid sampling points to restore the continuity of the rotation speed sequence, ultimately obtaining the corrected mechanical inertial response segment.

[0029] in, Indicates the first Local mean within the neighborhood window of each sampling point This represents the local standard deviation within the corresponding neighborhood window. The anomaly detection threshold is typically determined dynamically using the local mean ± the anomaly detection coefficient × the local standard deviation, where the anomaly detection coefficient... The threshold can be set to 2 to 4, with 3 being preferred. This is because when a sampling point deviates from the local mean by more than two standard deviations, it usually deviates significantly from the normal mechanical inertial fluctuation range. When it exceeds three standard deviations, most random noise, spike interference, and acquisition anomalies can be effectively identified, while avoiding excessive deletion of normal fluctuation data; therefore, a value of 3 is chosen. The corresponding anomaly judgment threshold dynamically changes with the intensity of local fluctuations, and its actual amplitude is generally controlled within the range of 1.5 to 4 times the amplitude of normal local fluctuations. Hydropower units inherently exhibit certain damped oscillation characteristics during the mechanical inertia recovery phase. If the threshold is set too small, normal mechanical fluctuations may be easily misjudged as anomalies; if the threshold is set too large, it may be unable to effectively eliminate abnormal amplitude points caused by sensor spike interference, electromagnetic pulses, or communication jitter, affecting the accuracy of subsequent fluctuation feature extraction.

[0030] S22: Extract the speed fluctuation feature set based on the corrected mechanical inertial response segment to characterize the dynamic fluctuation characteristics during the mechanical inertia recovery process of the hydropower unit. The speed stabilization duration feature characterizes the length of time it takes for the unit to recover to a stable operating state after the end of the regulation action; S221: Starting from the end of the adjustment action, the time difference between the moment when the speed fluctuation amplitude first enters the steady-state allowable band and remains stable and the moment when the fluctuation amplitude first enters the steady-state allowable band and remains stable is defined as the speed stabilization duration, expressed as: ; S221 begins by using the end of the adjustment action as a starting point. It continuously reads the rotational speed data from subsequent corrected mechanical inertial response segments and sets a steady-state allowable band, for example, centered on the estimated steady-state speed or steady-state mean, allowing fluctuations of 0.02% to 0.10% of the rated speed. Then, it sequentially checks whether the rotational speed fluctuations enter this allowable band, point by point or sliding window by sliding window. When, after a certain moment, the rotational speed fluctuations do not exceed the allowable band for multiple consecutive sampling periods or sliding windows, that moment is considered the first time the rotational speed has entered and remained stable. In other words, it's not enough for the rotational speed to momentarily fall into the allowable band; it requires continued observation for a period to confirm that the rotational speed does not significantly exceed the limit again. Finally, the time difference between this stable entry point and the end of the adjustment action is used as the characteristic of sustained rotational speed stabilization. The longer this characteristic is, the longer it takes for the unit to stabilize its mechanical speed after the adjustment is completed, which may reflect slow inertial response, insufficient damping, or abnormal mechanical transmission. The shorter this characteristic is and the more normal it is, the faster the unit's recovery process and the more normal its mechanical inertial response. Indicates the duration of the speed stabilization. Indicates the end time of the adjustment action. This indicates the moment when the rotational speed first enters and remains within the steady-state allowable band.

[0031] S222: The fluctuation decay characteristic characterizes the ability of rotational speed fluctuations to decrease over time. The peak envelope of the corrected mechanical inertial response segment is extracted, and the decreasing slope of the envelope over time is calculated. The fluctuation decay characteristic is expressed as: ; in, Indicates the rate of fluctuation decay. Indicates time The corresponding peak envelope amplitude, Indicates time The corresponding peak envelope amplitude.

[0032] The residual fluctuation characteristics near steady state characterize the intensity of high-frequency micro-oscillations that still exist after the unit enters a fully steady state; a preset analysis period is extracted after the fully steady state moment, and the residual fluctuation energy of the rotational speed relative to the steady-state mean is calculated; expressed as: ; in, This represents the residual fluctuation energy near the steady state. This represents the mean speed in steady state. This indicates the number of sampling points within the steady-state analysis period. Indicates the first The corrected rotational speed value corresponding to each sampling time. Indicates the sampling point number.

[0033] The asymmetric characteristic of positive and negative fluctuations characterizes the difference in energy distribution between positive and negative fluctuations in rotational speed. Specifically, the cumulative fluctuations of the rotational speed above and below the steady-state mean are statistically analyzed, and the degree of difference between the two is calculated, expressed as follows: ; in, This represents the asymmetric eigenvalues ​​of positive and negative fluctuations. This represents the cumulative amount of positive fluctuations. This represents the cumulative amount of negative fluctuations.

[0034] Local reverse fluctuation characteristics characterize local abnormal rebounds or reverse oscillations that occur during the overall fluctuation decay process. Based on the overall decay trend, fluctuation segments where the direction of local speed change is opposite to the main decay direction are detected, and the corresponding local reverse fluctuation amplitude is calculated; expressed as: ; in, Indicates the characteristic value of local reverse fluctuation. This indicates the sampling span corresponding to the local reverse fluctuation interval.

[0035] S3: Compare the speed fluctuation feature set with the normal speed fluctuation benchmark feature under the corresponding operating conditions of the hydropower unit to generate mechanical fluctuation deviation results. Identify the mechanical anomaly type based on the mechanical fluctuation deviation results, and generate the mechanical fault judgment result of the hydropower unit based on the mechanical anomaly type.

[0036] S3 specifically includes: S31: Based on the current operating condition information of the hydropower unit, retrieve the normal speed fluctuation benchmark feature set that matches the current operating state from the pre-established normal operating condition sample data; the operating condition information includes at least the unit load range, guide vane opening range, grid connection frequency status, and unit operating mode; the normal speed fluctuation benchmark feature set includes at least the normal speed stabilization persistence feature, normal fluctuation attenuation feature, residual fluctuation feature near normal steady state, normal positive and negative fluctuation asymmetry feature, and normal local reverse fluctuation feature. Subsequently, compare the speed fluctuation feature set extracted from the current corrected mechanical inertial response segment with the corresponding normal speed fluctuation benchmark features item by item, and calculate the degree of difference between each feature; The degree of difference characterizes the extent to which the current speed fluctuation characteristics deviate from the baseline characteristics under normal operating conditions; the degree of difference is expressed as: ; in, Indicates the first The degree of difference corresponding to class features This represents the first segment extracted from the current mechanical inertial response. Characteristic values ​​of speed fluctuation, This represents the baseline characteristic value corresponding to normal speed fluctuation.

[0037] When the dissimilarity corresponding to any feature satisfies: ; Then, it is determined that the corresponding feature has exceeded the normal fluctuation range, and the feature is marked as an abnormal fluctuation feature; where, Indicates the first The deviation tolerance thresholds corresponding to different characteristics can be set separately according to different speed fluctuation characteristics, and can be controlled within 10% to 35% of the corresponding normal reference characteristic value. Among them, the deviation tolerance range for speed stabilization and fluctuation decay characteristics is 15% to 25%, the deviation tolerance range for residual fluctuation characteristics near steady state and local reverse fluctuation characteristics is 20% to 35%, and the deviation tolerance range for positive and negative fluctuation asymmetry characteristics is 10% to 20%. Hydropower units will have certain normal fluctuation differences under different loads, heads, and operating conditions. Therefore, it is necessary to retain a reasonable tolerance range to avoid misjudging normal operating condition fluctuations as mechanical anomalies. Among them, speed stabilization and fluctuation decay characteristics are greatly affected by the unit's inertia and speed control system, and normal fluctuations are relatively stable. Therefore, the tolerance range can be appropriately reduced. Residual fluctuation characteristics near steady state and local reverse fluctuation characteristics are easily affected by hydraulic disturbances, measurement noise, and local transient oscillations. Therefore, the tolerance range needs to be appropriately increased. Positive and negative fluctuation asymmetry characteristics are more sensitive to rotor imbalance and asymmetry. To improve the sensitivity of anomaly identification, a smaller deviation tolerance range is adopted.

[0038] Subsequently, all abnormal fluctuation characteristics and their corresponding differences are summarized to form the mechanical fluctuation deviation results; the mechanical fluctuation deviation results include the type of abnormal fluctuation characteristic, the magnitude of the corresponding difference, and the number of abnormal characteristics.

[0039] S32: Identify the type of mechanical anomaly of the current hydropower unit based on the mechanical fluctuation deviation results; S321, input the abnormal fluctuation feature combination pattern and corresponding difference level in the mechanical fluctuation deviation results into the pre-established anomaly type discrimination rule base for matching analysis; the anomaly type discrimination rule base pre-stores the correspondence between different abnormal fluctuation feature combinations and mechanical anomaly types.

[0040] The exception type discrimination rule base mainly includes the following: (1) Set of mechanical anomaly types, used to store different mechanical fault categories, including at least rotational inertia anomaly, damping anomaly, misalignment or imbalance, and local friction or jamming; (2) Abnormal fluctuation feature combination rules, used to store the abnormal fluctuation feature combination relationship corresponding to different mechanical abnormality types, including the combination patterns between speed return to stabilization continuous feature abnormality, fluctuation attenuation feature abnormality, residual fluctuation feature abnormality near steady state, positive and negative fluctuation asymmetry feature abnormality and local reverse fluctuation feature abnormality; (3) Feature difference level rules, used to store the difference level classification standards corresponding to each abnormal fluctuation feature, including slight deviation, moderate deviation and severe deviation; (4) Feature weight information, used to characterize the importance of different abnormal fluctuation features in the corresponding mechanical anomaly identification process; (5) Fault severity mapping relationship, used to establish the correspondence between the difference in abnormal fluctuation characteristics and the severity level of mechanical faults; (6) Maintenance suggestion mapping information, used to establish maintenance suggestion labels corresponding to different mechanical abnormality types and severity.

[0041] The steps for building an exception type discrimination rule base include: First, historical rotational speed time series data of multiple hydropower units under normal operation and known mechanical abnormal conditions are collected, and the corresponding rotational speed fluctuation feature set is extracted according to the method of this invention. Among them, the known mechanical abnormal conditions include typical mechanical abnormal conditions such as changes in rotational inertia, damping degradation, shaft misalignment, rotor imbalance, and local friction jamming.

[0042] Subsequently, statistical analysis was conducted on the speed fluctuation characteristics corresponding to different mechanical abnormal states to obtain the characteristic change patterns under each abnormal state; for example, the growth trend of the duration of speed stabilization, the decline trend of the fluctuation decay rate, the change trend of the degree of asymmetry between positive and negative fluctuations, and the enhancement pattern of local reverse fluctuations were statistically analyzed.

[0043] Then, based on the characteristic change patterns of each mechanical abnormal state, a correspondence between abnormal fluctuation feature combinations and mechanical abnormality types is established; for the same mechanical abnormality type, its high-frequency abnormal feature combination patterns are statistically analyzed, and the corresponding difference level range and feature weights are determined.

[0044] Finally, the above-mentioned abnormality types, abnormal fluctuation feature combination rules, difference level rules, fault severity mapping relationship, and maintenance suggestion mapping relationship are uniformly stored to form an abnormality type discrimination rule library, which is then called and matched in the subsequent mechanical fault judgment process.

[0045] Table 1. Example Table of Anomaly Type Judgment Rule Base S322: When the speed stabilization characteristic is abnormal and the fluctuation decay characteristic is abnormal, it can be identified as an abnormal inertia type; when the fluctuation decay characteristic is abnormal and the residual fluctuation characteristic near the steady state is significantly increased, it can be identified as a damping abnormal type; when the asymmetric characteristic of positive and negative fluctuations is significantly abnormal, it can be identified as a misalignment or imbalance type; when the local reverse fluctuation characteristic is abnormally enhanced and accompanied by an increase in the residual fluctuation characteristic near the steady state, it can be identified as a local friction or jamming type. After completing the rule matching, the corresponding mechanical anomaly type result is output; if multiple anomaly type rules are satisfied simultaneously, the dominant mechanical anomaly type is determined according to the relative magnitude of the difference in the corresponding abnormal fluctuation characteristics.

[0046] S33: After identifying the type of mechanical anomaly, further combine the characteristic difference level in the mechanical fluctuation deviation results to conduct a comprehensive mechanical fault determination of the current hydropower unit. Based on the number of abnormal fluctuation characteristics, the magnitude of the difference corresponding to each characteristic, and the type of mechanical anomaly, a comprehensive assessment of the fault severity is conducted. The fault severity is expressed as follows: ; in, This indicates the severity rating of the fault. Indicates the first The weighting coefficients corresponding to the abnormal fluctuation characteristics Indicates the number of abnormal fluctuation characteristics. This indicates the degree of difference in the corresponding abnormal fluctuation characteristics.

[0047] Based on the range of the severity evaluation value, mechanical faults are classified into mild, moderate and severe abnormalities. At the same time, corresponding maintenance suggestion labels are output in combination with the type of mechanical abnormality. The maintenance suggestion labels include at least the types of continued operation and observation, planned maintenance, shutdown inspection and emergency maintenance, forming a complete mechanical fault judgment result of hydropower unit.

[0048] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments; however, those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0049] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for determining mechanical faults in hydropower units based on speed fluctuation characteristics, characterized in that, Includes the following steps: S1: During the operation of the hydropower unit, the unit's operating status data is acquired in real time, the end time of the adjustment action is identified, and the speed time sequence data segment between the end time of the adjustment action and the moment when the unit enters a completely steady state is extracted, using the end time of the adjustment action as the starting point, to obtain the mechanical inertial response segment. S2: Perform benchmark steady-state correction and disturbance removal processing on the mechanical inertial response segment to obtain the corrected mechanical inertial response segment; extract the rotational speed fluctuation feature set based on the corrected mechanical inertial response segment; S3: Compare the speed fluctuation feature set with the normal speed fluctuation benchmark feature under the corresponding operating conditions of the hydropower unit to generate mechanical fluctuation deviation results, identify the mechanical anomaly type based on the mechanical fluctuation deviation results, and generate the mechanical fault judgment result of the hydropower unit in combination with the mechanical anomaly type.

2. The method for determining mechanical faults in hydropower units based on speed fluctuation characteristics according to claim 1, characterized in that, S1 specifically includes: S11: Synchronously acquire unit operating status data at a sampling frequency capable of completely capturing transient changes in rotational speed, and store the acquired signals in time alignment; the unit operating status data includes rotational speed signal, guide vane opening command signal, grid connection frequency signal, and load status signal; S12: Real-time monitoring of the rate of change of the guide vane opening command signal, the grid connection frequency signal, and the load status signal to determine the end time of the adjustment action; S13: Taking the end time of the adjustment action as the starting point of the interception, the sliding window fluctuation amplitude of the speed signal relative to the estimated steady-state speed is calculated in real time. When the sliding window fluctuation amplitude is less than the preset speed steady-state determination threshold in multiple consecutive sliding windows, the unit is determined to have entered the complete steady-state moment. The speed time sequence data segment from the interception starting point to the complete steady-state moment is extracted as the mechanical inertial response segment.

3. The method for determining mechanical faults in hydropower units based on speed fluctuation characteristics according to claim 3, characterized in that, The speed signal is used to characterize the real-time speed change of the rotating components of the unit, the guide vane opening command signal is used to characterize the change of the control command of the speed regulation system on the guide vane opening, the grid connection frequency signal is used to characterize the frequency stability of the power grid currently connected to the unit, and the load status signal is used to characterize whether the unit load has changed or been disturbed.

4. The method for determining mechanical faults in hydropower units based on speed fluctuation characteristics according to claim 2, characterized in that, The conditions for determining the end time of the adjustment action specifically include: Condition 1: The rate of change is consistently less than the pre-calibrated governor command stability threshold and the guide vane opening command signal does not exceed the adjustment dead zone change in subsequent consecutive sampling periods; Condition 2: The deviation between the grid connection frequency signal and the rated frequency is continuously less than the preset frequency tolerance threshold, and the load status signal remains unchanged for multiple consecutive sampling periods; If both conditions one and two are met, then the current time will be determined as the end time of the adjustment action.

5. The method for determining mechanical faults in hydropower units based on speed fluctuation characteristics according to claim 1, characterized in that, S2 specifically includes: S21: Perform a reference steady-state correction on the rotational speed time series data in the mechanical inertial response segment. The reference steady-state correction is to fit the overall rotational speed change trend in the mechanical inertial response segment and subtract the overall change trend from the original rotational speed signal to obtain a correction signal after eliminating the reference offset. Perform disturbance removal processing on the correction signal to obtain the corrected mechanical inertial response segment. S22: Extract a speed fluctuation feature set based on the corrected mechanical inertial response segment. The speed fluctuation feature set includes speed stabilization and persistence features, fluctuation attenuation features, residual fluctuation features near steady state, positive and negative fluctuation asymmetry features, and local reverse fluctuation features.

6. The method for determining mechanical faults in hydropower units based on speed fluctuation characteristics according to claim 5, characterized in that, The disturbance removal process includes identifying and removing abnormal amplitude points caused by instantaneous sensor interference or external random impact, and interpolating and repairing the removed points to obtain the corrected mechanical inertial response segment.

7. The method for determining mechanical faults in hydropower units based on speed fluctuation characteristics according to claim 5, characterized in that, The sustained characteristic of speed stabilization includes the length of time from the end of the adjustment action until the speed fluctuation amplitude first enters and remains within the steady-state allowable band; the fluctuation attenuation characteristic includes the attenuation rate of the peak and trough values ​​of speed fluctuation over time or the slope of the envelope descent in the corrected mechanical inertial response segment; the residual fluctuation characteristic near the steady state includes the high-frequency residual fluctuation amplitude or energy of the speed relative to the steady-state mean over a period of time after the unit enters a fully steady state; the asymmetric positive and negative fluctuation characteristic includes the degree of difference between the cumulative amount of positive and negative speed fluctuation amplitudes in the mechanical inertial response segment; the local reverse fluctuation characteristic includes local speed recovery or oscillation amplitudes that occur opposite to the main attenuation direction under the overall attenuation trend.

8. The method for determining mechanical faults in hydropower units based on speed fluctuation characteristics according to claim 1, characterized in that, S3 specifically includes: S31: Obtain a normal speed fluctuation benchmark feature set that matches the current operating conditions of the hydropower unit. Compare each feature in the speed fluctuation feature set with the corresponding normal speed fluctuation benchmark feature one by one, calculate the difference between each feature, and when the difference of any feature exceeds the corresponding preset deviation tolerance range, mark the corresponding feature as an abnormal fluctuation feature. Collect all abnormal fluctuation features and differences to generate the mechanical fluctuation deviation result. S32: Identify the mechanical anomaly type based on the mechanical fluctuation deviation result, and map it to the pre-established anomaly type discrimination rule base according to the combination pattern of different abnormal fluctuation characteristics and the relative magnitude of the difference. S33: Combining the mechanical anomaly type and the characteristic difference level in the mechanical fluctuation deviation result, a comprehensive mechanical fault judgment result for the hydropower unit is generated. The mechanical fault judgment result includes the fault type, fault severity level, and corresponding maintenance suggestion label.

9. The method for determining mechanical faults in hydropower units based on speed fluctuation characteristics according to claim 8, characterized in that, The normal speed fluctuation reference feature set includes at least the normal speed stabilization and persistence feature, the normal fluctuation attenuation feature, the residual fluctuation feature near the normal steady state feature, the normal positive and negative fluctuation asymmetry feature, and the normal local reverse fluctuation feature.

10. The method for determining mechanical faults in hydropower units based on speed fluctuation characteristics according to claim 8, characterized in that, The anomaly type discrimination rule base includes the correspondence between feature combinations and mechanical anomaly types, which include rotational inertia anomalies, damping anomalies, misalignment or imbalance, and local friction or jamming.

Images