Method and system for evaluating degradation degree of wind power gear box

Abstract

This invention belongs to the field of wind turbine gearbox fault diagnosis technology. Addressing the problems of existing assessment methods that rely solely on amplitude statistical characteristics, are susceptible to noise interference, and depend on complex model training, this invention proposes a method and system for assessing the degradation degree of wind turbine gearboxes. It performs structured analysis of vibration signals through zero-crossing interval division, extracting interval extrema as impact intensity features; constructs amplitude reference thresholds using healthy state samples, and enhances the characterization of abnormal impacts through multi-level amplitude coding; furthermore, it identifies periodic impact structures in the vibration signals through autocorrelation analysis and calculates the periodic intensity of intervals; and introduces periodic weights on the traditional Gini index to construct a periodically perceived weighted Gini index, achieving a quantitative assessment of the degradation degree of wind turbine gearboxes, which is beneficial for condition monitoring and life prediction.

Description

Technical Field

[0001] This invention relates to the field of wind turbine gearbox fault diagnosis technology, specifically to a method and system for assessing the degree of degradation of wind turbine gearboxes. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] The wind turbine gearbox is a core component of the wind turbine's transmission system. Operating under constant random wind and impact loads, its internal gears and bearings are highly susceptible to progressive degradation failures such as wear and pitting. To ensure safe turbine operation, vibration signal analysis-based condition monitoring technology has become mainstream in the industry. Existing technologies typically quantify the equipment's degradation state by collecting gearbox vibration signals and constructing health indicators. With the development of rotating machinery condition monitoring technology, the key to achieving full lifecycle management of wind turbine gearboxes lies in extracting subtle fault characteristics from complex background noise and transforming them into continuous and stable degradation indicators.

[0004] However, existing degradation assessment methods struggle to comprehensively characterize both impact intensity and periodic structure simultaneously. Current techniques largely rely on traditional amplitude statistics or unweighted Gini coefficients, often focusing solely on signal amplitude distribution while neglecting the periodic clustering of impact events over time during fault evolution. Specifically, existing methods lack an effective mechanism to deeply integrate the "periodic intensity" and "amplitude distribution" of vibration signals. This prevents the allocation of higher weights to periodic impacts with fault diagnosis significance when calculating degradation indices. This technical deficiency makes it difficult for existing indices to accurately distinguish between random impacts and periodic impacts caused by faults in complex noise environments, leading to insensitive degradation trend responses or misjudgments, and failing to meet the demands of online monitoring for high robustness and absolute quantitative assessment. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method and system for assessing the degradation degree of wind turbine gearboxes. Without relying on complex model training, it can simultaneously utilize vibration and shock intensity and periodic structural information to achieve stable characterization of the degradation process of wind turbine gearboxes. This method is suitable for online condition monitoring and degradation trend assessment of wind turbine units.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] In a first aspect, the present invention provides a method for assessing the degradation degree of wind turbine gearboxes.

[0008] A method for assessing the degradation level of wind turbine gearboxes includes the following steps:

[0009] The original vibration signal of the wind turbine gearbox is acquired, the original vibration signal is preprocessed, and the preprocessed vibration signal is divided into multiple vibration intervals based on the zero crossover point. The extreme value amplitude of each vibration interval is extracted to obtain the interval extreme value sequence.

[0010] An amplitude reference threshold is constructed based on historical vibration samples under healthy conditions. The interval extreme value sequence is then hierarchically mapped according to the amplitude reference threshold to generate an interval amplitude intensity coding sequence.

[0011] Autocorrelation analysis is performed on the original vibration signal, and the delay characteristics in the impact anomaly sequence are identified based on the autocorrelation analysis results to obtain the periodic intensity characteristics of the vibration interval.

[0012] The interval amplitude intensity coding sequence and periodic intensity features are simultaneously sorted to obtain an ordered periodic intensity sequence. Periodic weights are constructed based on the ordered periodic intensity sequence, and the Gini index is weighted and corrected based on the periodic weights to generate a periodic perception weighted Gini index.

[0013] Based on the repeated calculation of the periodic perception-weighted Gini index using a sliding time window, a degradation index sequence is generated to characterize the degradation evolution process of wind turbine gearboxes.

[0014] In one implementation of the first aspect of the present invention, an amplitude reference threshold is constructed based on historical vibration samples under healthy conditions, and a hierarchical mapping is performed on the interval extreme value sequence according to the amplitude reference threshold to generate an interval amplitude intensity encoded sequence, including:

[0015] Calculate the high quantiles of the extreme value sequences of multiple historical health status samples, and fuse the high quantiles to obtain the health reference threshold;

[0016] Each extreme value in the interval extreme value sequence is compared with a health reference threshold, and the extreme values ​​are mapped to different intensity levels based on the comparison results to generate an interval amplitude intensity coding sequence.

[0017] In one implementation of the first aspect of the present invention, mapping extreme values ​​to different intensity levels based on comparison results to generate an interval amplitude intensity coding sequence includes:

[0018] The extreme value of the interval is compared with the health reference threshold. If the extreme value of the interval is less than or equal to the health reference threshold, the first intensity level code is assigned.

[0019] If the extreme value of the interval is greater than the health reference threshold, then calculate the multiple by which the extreme value of the interval exceeds the health reference threshold;

[0020] Based on the preset multiple interval division rules, different multiple intervals are mapped to different intensity level codes, generating interval amplitude intensity code sequences.

[0021] In one implementation of the first aspect of the present invention, autocorrelation analysis is performed on the original vibration signal, and delay features in the impact anomaly sequence are identified based on the autocorrelation analysis results to obtain the periodic intensity features of the vibration interval, including:

[0022] Anomaly thresholds are constructed based on health reference thresholds and absolute median difference methods. Based on the anomaly thresholds, interval amplitude intensity encoded sequences are detected to generate shock anomaly sequences.

[0023] Calculate the autocorrelation function of the shock anomaly sequence, and estimate the shock period based on the location of the maximum value of the autocorrelation function;

[0024] Based on the impact period, the consistency of the impact anomaly sequence is calculated in the neighborhood, and the periodic intensity characteristics of the vibration interval are obtained.

[0025] In one implementation of the first aspect of the present invention, the interval amplitude intensity coding sequence and the periodic intensity feature are simultaneously sorted to obtain an ordered periodic intensity sequence. Periodic weights are constructed based on the ordered periodic intensity sequence, and the Gini index is weighted and corrected based on the periodic weights to generate a periodically perceived weighted Gini index, including:

[0026] The interval amplitude intensity coding sequence is sorted in ascending order to obtain an ordered amplitude sequence. Based on the mapping relationship between the interval amplitude intensity coding sequence and the ordered amplitude sequence, the periodic intensity features are rearranged to obtain an ordered periodic intensity sequence.

[0027] The weight coefficients for each interval are calculated based on the ordered periodic intensity sequence to obtain the periodic weight sequence;

[0028] Based on the periodic weight sequence and the ordered amplitude sequence, the periodic-perceived weighted Gini index is calculated through weighted summation and accumulation operations. In one implementation of the first aspect of this invention, the calculation of the periodic-perceived weighted Gini index based on the periodic weight sequence and the ordered amplitude sequence through weighted summation and accumulation operations includes:

[0029] To calculate the weighted cumulative sum, multiply each element in the ordered magnitude sequence by its corresponding period weight, and accumulate all the product results from the first element to the current element.

[0030] Calculate the weighted sum of the total energy, and sum the products of all ordered magnitudes and their corresponding period weights;

[0031] Construct the weighted area under the Lorentz curve based on the ratio of the weighted cumulative sum to the weighted sum of total energy.

[0032] The periodic perceived weighted Gini index is obtained by calculating the non-uniformity of the numerical distribution based on the weighted area.

[0033] In one implementation of the first aspect of the present invention, the logic of the step of generating an interval amplitude intensity encoded sequence by performing hierarchical mapping on the interval extreme value sequence according to the amplitude reference threshold is as follows:

[0034] When the extreme value of the interval is less than or equal to the health reference threshold, the output code is 0;

[0035] When the extreme value of the interval is greater than the health reference threshold but less than or equal to twice the health reference threshold, output code 1;

[0036] When the extreme value of the interval is greater than twice the health reference threshold and less than or equal to three times the health reference threshold, output code 2;

[0037] When the extreme value of the interval is greater than 3 times the health reference threshold and less than or equal to 4 times the health reference threshold, output code 3;

[0038] When the extreme value of the interval is greater than 4 times the health reference threshold, the output code is 4;

[0039] Arrange the above codes in chronological order to generate an interval amplitude intensity coding sequence.

[0040] Secondly, the present invention provides a system for assessing the degradation level of wind turbine gearboxes.

[0041] A wind turbine gearbox degradation assessment system includes:

[0042] The signal processing unit is configured to: acquire the original vibration signal of the wind turbine gearbox, preprocess the original vibration signal, divide the preprocessed vibration signal into multiple vibration intervals based on the zero crossover point, extract the extreme amplitude of each vibration interval, and obtain the interval extreme value sequence.

[0043] The amplitude encoding unit is configured to: construct an amplitude reference threshold based on historical vibration samples under healthy conditions, perform hierarchical mapping on the interval extreme value sequence according to the amplitude reference threshold, and generate an interval amplitude intensity encoding sequence;

[0044] The periodic quantization unit is configured to: perform autocorrelation analysis on the original vibration signal, identify the delay characteristics in the impact anomaly sequence based on the autocorrelation analysis results, and obtain the periodic intensity characteristics of the vibration interval;

[0045] The index calculation unit is configured to: synchronously sort the interval amplitude intensity coding sequence and periodic intensity features to obtain an ordered periodic intensity sequence; construct periodic weights based on the ordered periodic intensity sequence; and perform weighted correction on the Gini index based on the periodic weights to generate a periodic perception weighted Gini index.

[0046] The trend building unit is configured to repeatedly calculate the periodically perceived weighted Gini index based on a sliding time window to generate a degradation index sequence to characterize the degradation evolution process of wind turbine gearboxes.

[0047] Thirdly, the present invention provides a computer-readable storage medium storing a computer program adapted to be loaded by a processor and executed by the wind turbine gearbox degradation assessment method of the first aspect of the present invention.

[0048] Fourthly, the present invention provides a computer device, comprising: a processor and a computer-readable storage medium;

[0049] A processor, adapted to execute computer programs;

[0050] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the wind turbine gearbox degradation assessment method of the first aspect of the present invention.

[0051] Compared with the prior art, the beneficial effects of the present invention are:

[0052] This invention effectively filters high-frequency random fluctuations in signals while preserving physically meaningful impact energy characteristics through a preprocessing method of "zero-crossing interval division + interval extreme value extraction." Based on this, it innovatively proposes a computational framework of "periodic perception weighted Gini index": First, it uses a health reference threshold to hierarchically map interval extreme values, generating amplitude intensity codes to enhance the characterization ability of abnormal impacts; then, it identifies the delay characteristics of the impact anomaly sequence through autocorrelation analysis, quantifying the periodic intensity of the vibration interval. Finally, it transforms the periodic intensity characteristics into weight coefficients to weight and correct the traditional Gini index. This mechanism assigns higher weights to "periodic impacts" with fault diagnosis significance during index calculation, while random impacts, lacking periodicity, are assigned lower weights, thus accurately distinguishing fault characteristics from interference signals under strong background noise; by introducing periodic weights to correct the Gini index, this method maps the degradation degree of wind turbine gearboxes to an absolute value range of [0,1]. Compared to traditional kurtosis or unweighted Gini index, which only reflects relative fluctuations, the degradation index sequence generated by this invention can more smoothly and stably track the entire life cycle evolution of equipment from health to failure. This invention does not rely on complex deep learning model training, has low computational complexity, and effectively overcomes the problems of existing statistical indicators being sensitive to noise, having large trend fluctuations, and being difficult to quantify absolutely. It is particularly suitable for online condition monitoring and life prediction of wind turbine units under complex operating conditions.

[0053] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0054] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0055] Figure 1 A flowchart illustrating a method for assessing the degradation level of a wind turbine gearbox, provided as an exemplary embodiment of the present invention;

[0056] Figure 2 A schematic diagram of different degradation stage encoding sequences is provided as an exemplary embodiment of the present invention, wherein, Figure 2 In the table, (a) represents the coded value for the health stage. Figure 2 (b) in the table represents the coded value for the early failure stage. Figure 2 In the equation (c), the coded value represents the intermediate fault stage. Figure 2 In this context, (d) represents the coded value for the failure stage;

[0057] Figure 3 A schematic diagram showing a trend comparison of different degradation assessment methods provided as an exemplary embodiment of the present invention, wherein, Figure 3 (a) in the text is used for A schematic diagram for assessing the bearing degradation process. Figure 3 (b) in the diagram is a schematic diagram of using the original Gini index to assess the bearing degradation process. Figure 3 (c) in the diagram is a schematic diagram of using kurtosis to assess the bearing degradation process;

[0058] Figure 4 A schematic diagram of a wind turbine gearbox degradation assessment system provided as an exemplary embodiment of the present invention;

[0059] Figure 5 A schematic diagram of a computer device provided for an exemplary embodiment of the present invention. Detailed Implementation

[0060] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0061] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0062] The gearbox is a critical component in the transmission system of a wind turbine, and its operating status directly affects the reliability and operational safety of the wind turbine. During the development of faults such as gear wear, tooth surface spalling, or bearing damage, vibration signals typically exhibit characteristics such as increased impact components, a more pronounced periodic structure, and a gradual concentration of energy distribution. However, in the complex operating environment of wind turbines, vibration signals are often simultaneously affected by multiple factors, including aerodynamic disturbances, structural vibrations, and background noise. This results in relatively weak early degradation characteristics, making it difficult for traditional vibration statistical indicators to accurately reflect the gearbox's degradation state.

[0063] The existing wind turbine gearbox condition assessment methods mainly have the following technical problems: (1) They rely solely on amplitude statistical features, making it difficult to characterize periodic impact characteristics. Some degradation assessment methods mainly construct health indicators based on amplitude statistical indicators such as root mean square value and kurtosis. However, these indicators mainly reflect changes in signal amplitude and are not sensitive enough to the periodic impact characteristics caused by faults such as gear meshing and bearing collisions. They are easily affected by random vibration in complex noise environments. (2) Single-point amplitude analysis is easily affected by noise. Existing methods usually directly perform statistical analysis on vibration signal sampling points. However, there are strong random fluctuations in the vibration signals of wind turbine units. Single-point amplitude changes may be caused by noise, resulting in large fluctuations in degradation indicators and reducing the stability of the assessment. (3) Some methods rely on complex model training. In recent years, some deep learning-based condition assessment methods require a large amount of historical data for model training and rely on complex network structures for feature learning. In the actual wind farm operation environment, due to the unstable quality of monitoring data or the lack of complete degradation data, model training and application are difficult.

[0064] In view of the problems existing in the existing solutions, this invention proposes a method for assessing the degradation degree of wind turbine gearboxes. First, the vibration signal of the wind turbine gearbox is preprocessed, and the vibration signal is structurally decomposed by dividing it into zero-crossing intervals, extracting the extreme values ​​of the vibration intervals as impact intensity features. Then, an amplitude reference threshold is constructed using healthy state samples, and the expression of abnormal impact features is enhanced through a multi-level amplitude encoding method. Further, autocorrelation analysis is used to identify periodic impact structures in the vibration signal, and the periodic intensity of the vibration interval is calculated. Based on this, a periodic weight is introduced to improve the traditional Gini index, constructing a periodically sensed weighted Gini index to achieve a quantitative expression of the concentration of vibration impact energy. Finally, an index sequence is calculated through a sliding time window to achieve continuous tracking and assessment of the wind turbine gearbox degradation process.

[0065] More specifically, the present invention comprises five steps, such as... Figure 1 As shown, it specifically includes:

[0066] S1: Preprocessing of vibration signals from wind turbine gearboxes and construction of interval structures;

[0067] S2: Interval amplitude modeling and intensity coding;

[0068] S3: Periodic shock identification and periodic quantification;

[0069] S4: Periodic-perceived weighted Gini index;

[0070] S5: Construction of degradation index sequence.

[0071] Step S1 of this implementation is used to extract the vibration interval with physical structural significance from the vibration signal of the wind turbine gearbox. Since gear meshing impact usually exhibits a periodic vibration response, the zero-crossing interval can better characterize the vibration half-cycle structure.

[0072] S101: Vibration signal acquisition.

[0073] Considering that the vibration signal of the wind turbine gearbox is a continuous-time signal, it is necessary to obtain a discrete vibration sequence through sensor sampling. Therefore, the discrete representation of the vibration signal is defined as follows: ;

[0074] in: For the first The vibration amplitude of the wind turbine gearbox at each sampling point; This represents the number of sampling points within a single time window.

[0075] S102: Signal mean removal processing.

[0076] Considering that the vibration signal may contain a DC component caused by sensor bias or low-frequency structural vibration, the vibration signal is subjected to mean-reduction processing to eliminate this effect. The calculation formula is as follows:

[0077] (1);

[0078] in, The vibration signal after removing the mean; This represents the average value of the vibration signal within the current time window. Formula (1) eliminates the DC component caused by sensor bias and low-frequency vibration, avoiding interference from the DC component in the extraction of vibration impact features and ensuring the accuracy of subsequent interval division and extreme value calculation.

[0079] S103: Zero-crossing interval division.

[0080] Considering that gear meshing impact will cause periodic oscillations in the vibration signal, in order to identify the half-periodic structure of the vibration waveform, the symbol sequence of the signal is first calculated:

[0081] (2);

[0082] in: It is a sequence of symbols; The sign function is used. By using formula (2), the vibration signal is converted into a sequence of positive and negative signs, which quickly locates the trend of signal amplitude change, provides a clear basis for zero-crossing point detection, and simplifies the half-cycle structure identification process.

[0083] Further detection of sign change locations is needed to determine the set of zero-crossing points of the vibration signal:

[0084] (3);

[0085] in: The set of zero-intersection locations; For the first The sampling locations of the zero-crossing points; The zero-crossing point number; Representing the The sign of the sampling position where each zero-crossing point is located; Representing the The sign of the sampling position where the zero crossover point is located. Through formula (3), the positive and negative abrupt change position of the vibration signal is accurately identified, the half-cycle boundary point is determined, and the continuous signal is structurally segmented to adapt to the periodic characteristics of gear fault impact.

[0086] Based on the zero-crossing point location, the vibration signal is divided into several vibration intervals, which are defined as follows:

[0087] (4)

[0088] in, For the first One vibration interval; The vibration interval number; Representing the Zero intersection points; Representing the There are zero-crossing points. By using formula (4), the half-cycle interval is divided with the zero-crossing points as the boundary, which fully preserves the impact energy of a single vibration and avoids the evaluation error caused by single-point sampling fluctuations.

[0089] Step S2 of this implementation is used to describe the vibration and impact intensity of the wind turbine gearbox and to establish a reference threshold through health status data.

[0090] S201: Interval extreme value extraction.

[0091] Considering that gear failures typically manifest as instantaneous impact events, with their energy usually peaking within the vibration range, the maximum amplitude is extracted for each vibration range as the impact intensity characteristic. The calculation formula is as follows:

[0092] (5);

[0093] in: For the first The extreme amplitude of each vibration interval; This is for absolute value operations; Calculate the maximum value; Representing the Each vibration interval is analyzed. The maximum amplitude of a single vibration interval is extracted using formula (5), highlighting the instantaneous impact intensity caused by the fault, filtering out high-frequency random noise, and retaining the core fault characteristic information.

[0094] This yields the interval extreme value sequence within the time window:

[0095] (6);

[0096] in: This represents the number of vibration intervals within the time window, where... These represent the extreme amplitudes of the 1st, 2nd to Kth vibration intervals, respectively.

[0097] S202: Construction of health reference thresholds.

[0098] Considering that the vibration amplitude of a wind turbine gearbox in a healthy state has a stable statistical distribution, a reference threshold is established using historical samples from the healthy state. First, the high quantile of the extreme value sequence of the healthy samples is calculated, and its expression is:

[0099] (7);

[0100] in: For the first The amplitude threshold for each healthy sample; 99th percentile calculation; For the first The extreme value sequence of a healthy sample. By using formula (7), the 99th percentile of the extreme value sequence of the healthy state is selected to lock the upper limit of the health amplitude, eliminate the interference of extreme outliers, and construct a robust health benchmark.

[0101] Subsequently, the threshold values ​​of multiple healthy samples are fused to obtain a more stable reference threshold, calculated using the following formula:

[0102] (8);

[0103] in: This is the fused health reference threshold; Number of healthy samples; This represents the amplitude threshold for the 1st, 2nd to Mth healthy samples. By using formula (8), the thresholds of multiple healthy samples are merged and the minimum value is taken, which improves the conservatism of the thresholds and avoids the early weak fault impacts being missed.

[0104] S203: Multi-level amplitude encoding.

[0105] Considering that a wind turbine gearbox failure can lead to a gradual increase in vibration and impact amplitude, the extreme values ​​within a range are mapped to multiple intensity levels to enhance the abnormal impact characteristics. The coding rule is as follows:

[0106] (9);

[0107] in, For interval The corresponding amplitude intensity level.

[0108] This yields the amplitude encoding sequence:

[0109] (10);

[0110] in, These represent the amplitude intensity levels of the 1st, 2nd to Kth vibration intervals, respectively.

[0111] like Figure 2 As shown, Figure 2 (a) Figure 2 (b) Figure 2 (c) and Figure 2 (d) in the figure shows the coding values ​​for the health stage, early failure stage, mid-failure stage and failure stage respectively. It can be seen that only a small number of 1s are generated in the early stage; after the failure occurs, the occurrence of 1s will increase rapidly; as the failure gradually becomes more serious, more 2, 3 and 4 codes will appear.

[0112] In step S3 of this implementation, gear failures in wind turbine gearboxes typically manifest as periodic impacts related to the gear meshing frequency, thus requiring identification of the impact cycle.

[0113] S301: Shock anomaly sequence.

[0114] Considering that the amplitude of abnormal impacts is usually higher than that of background vibrations, an impact anomaly sequence is constructed by threshold detection, and its expression is:

[0115] (11);

[0116] in: For shock anomaly indicator variables; For detection features; The threshold value is used to filter out abnormal amplitude ranges using formula (11), mark potential fault impact locations, and effectively distinguish between normal vibration and fault abnormal signals.

[0117] Abnormal threshold Calculated using the absolute median method:

[0118] (12);

[0119] in: This is the threshold adjustment coefficient; It is the absolute median difference function; The median is a function of the median. Using formula (12), the adaptive threshold is calculated using the absolute median difference to enhance the ability to resist noise interference and adapt to the complex and ever-changing operating conditions of wind farms.

[0120] S302: Period estimation.

[0121] Considering that periodic shocks can produce recurring patterns in anomalous sequences, the shock period is estimated using the autocorrelation function, and its calculation formula is as follows:

[0122] (13);

[0123] in, It is the autocorrelation function; This is the delay amount; Represents an indicator variable for shock anomalies; Represents delay Abnormal variables; Represents the total number of vibration intervals; This represents the delay amount. By using formula (13), the delay correlation of abnormal sequences is calculated, and the recurring impact pattern is captured, providing a reliable basis for accurate estimation of the fault impact cycle.

[0124] Then the impact period for:

[0125] (14);

[0126] in, Represents the autocorrelation function Delay amount of the maximum value By using formula (14), the maximum delay position of the autocorrelation function can be located, the fault impact cycle can be accurately identified, and the gear meshing frequency correlation characteristics can be matched.

[0127] S303: Periodic intensity.

[0128] Considering that periodic shocks exhibit similar characteristics at multiple periodic locations, periodic consistency is calculated using neighborhood periodic locations, and its expression is as follows:

[0129] (15);

[0130] in, The intensity is periodic within the interval; Number of search cycles; Represents moving forward Anomaly marker values ​​for each period; Represents backward Anomaly marker values ​​for each period; The vibration interval number is represented by formula (15). The consistency of outliers in the multi-period neighborhood is calculated, the periodic intensity of the interval is quantified, and the periodic impact of the fault period is effectively distinguished from random noise interference.

[0131] In step S4 of this implementation method, vibration amplitude information and periodic impact information are integrated to construct a wind turbine gearbox degradation index.

[0132] S401: The amplitude coding sequence is synchronously sorted with the interval periodic intensity.

[0133] Considering that the Gini index is used to measure the non-uniformity of numerical distribution, the amplitude encoding sequence of formula (10) is first performed. Sort in ascending order to obtain an ordered amplitude sequence:

[0134] (16);

[0135] Suppose the above sorting mapping relationship for:

[0136] (17);

[0137] While sorting, the same mapping relationship is used. For interval periodicity intensity Rearrangement yields an ordered periodic intensity sequence:

[0138] (18);

[0139] S402: Periodic weights.

[0140] Considering that periodic shocks are more likely to be caused by gear failures, a periodic weighting coefficient is introduced into the ordered periodic intensity sequence. Its expression is:

[0141] (19);

[0142] in, This is the weighting coefficient, which is set to 2 here; Representing the An ordered periodic intensity; Representing the An ordered periodic intensity is used. Through formula (19), a weighting coefficient is constructed based on the periodic intensity to assign higher weights to periodic fault shocks, reduce the impact of random shocks, and improve the robustness of the index.

[0143] The periodic weight sequence is obtained:

[0144] (20)

[0145] S403: Calculate the weighted Gini index.

[0146] After introducing periodic weights, the periodicity-perceived weighted Gini index is calculated as follows:

[0147] (twenty one);

[0148] in, This is an indicator of the degree of degradation of wind turbine gearboxes. According to... Figure 2 The encoded values ​​at different degradation stages are shown, revealing that only a small number of 1s are generated in the early stages. 0; After a fault occurs, the occurrence of 1s will increase rapidly. Initially greater than 0.1; as the fault worsens, more codes 2, 3, and 4 appear. The value gradually approaches 1. By using formula (21), the Gini index is corrected by integrating the amplitude and periodic characteristics, and the degree of degradation is mapped to the [0,1] interval, achieving absolute quantification and a smooth and stable trend.

[0149] In step S5 of this implementation, considering that the degradation of wind turbine gearboxes is a process that evolves over time, a sliding time window is used to repeatedly calculate the above indicators to obtain the degradation sequence:

[0150] (twenty two);

[0151] in: There are W samples; This sequence represents the total number of lifetime samples and is used to characterize the degradation and evolution process of wind turbine gearboxes. This is the first sample. This is the second sample; These represent the degradation indices for the first sample, the second sample, and so on up to the Wth sample.

[0152] Figure 3 Demonstrating the evaluation method proposed in this invention A schematic diagram showing the original Gini index and kurtosis used to assess the bearing degradation process. (From...) Figure 3As shown in (a), the evaluation index constructed in this invention exhibits a gradual upward trend over time, showing a significant increase after approximately 300 minutes, and maintaining a relatively stable upward trend in the later stages, effectively reflecting the entire process of the bearing's evolution from a healthy state to a failure state. The index changes relatively slowly in the early stages of degradation, but shows a significant response during the impact enhancement stage, indicating that it can effectively characterize the key stages of the degradation process. Furthermore, the evaluation index constructed in this invention is generally within the range of [0, 1], achieving an absolute measurement of the degree of degradation.

[0153] Depend on Figure 3 As shown in (b), although the original Gini index can reflect a certain degree of change trend during the degradation process, the overall fluctuation is quite obvious, especially in the middle and later stages where there are multiple local fluctuations. This is mainly because the original Gini index only reflects the unevenness of the vibration amplitude distribution and is more sensitive to random shocks and noise disturbances, which affects the stability of the index.

[0154] Depend on Figure 3 As shown in (c), the kurtosis index changes little over most of the time period, only abruptly changing when the impact intensifies significantly in the later stages, demonstrating its insensitivity to the early degradation phase. Furthermore, the kurtosis fluctuates considerably in the later stages, making it difficult to establish a continuous and stable degradation trend.

[0155] comprehensive Figure 3 The comparison shows that the method of this invention incorporates periodic impact information when constructing the index, enabling the index to not only reflect the vibration energy distribution characteristics but also the periodic structure of the impact, thus exhibiting a more continuous and stable trend during degradation. Compared to the original Gini index and kurtosis index, which are based solely on amplitude statistical characteristics, this method improves the overall characterization ability of the degradation process while suppressing the influence of random fluctuations. Therefore, the evaluation method proposed in this invention can more stably describe the bearing degradation evolution process, which is beneficial for subsequent condition monitoring and degradation trend analysis.

[0156] In summary, this invention achieves an absolute measurement of the degradation degree of wind turbine gearbox components. By constructing a periodically perceptually weighted Gini index, this invention maps the concentration of vibration and impact energy in wind turbine gearboxes into a continuous degradation index, mapping the degradation degree to the interval [0,1]. A value close to 1 indicates highly concentrated impact energy and a high degree of failure, while a value close to 0 indicates relatively uniform energy distribution and healthy bearings. Unlike traditional statistical indicators such as kurtosis, mean, and RMS, which rely on sample distribution and amplitude fluctuations and only reflect relative trends, this method achieves an absolute quantitative expression of the degradation degree of gear and bearing failure.

[0157] This invention effectively identifies the periodic impact characteristics generated by gear faults. Gear faults in wind turbine gearboxes typically produce periodic impact vibrations related to the gear meshing frequency. However, in complex operating environments, a large amount of random noise and background vibration can easily mask these periodic impacts, making it difficult for traditional statistical indicators to accurately identify fault characteristics. This method identifies the periodic impact structure in the vibration signal and incorporates the periodic information into the degradation index calculation, giving the periodic impact caused by gear damage a higher weight in the index calculation, thereby improving the ability to identify gear faults. Therefore, this method can more accurately reflect the impact characteristics caused by gear damage in complex vibration environments.

[0158] This invention reduces the impact of random noise on degradation assessment results. Wind turbine operating environments are complex, and vibration signals typically contain various interferences such as aerodynamic disturbances, structural vibrations, and motor noise. These random factors can easily cause significant fluctuations in traditional vibration indices, thus affecting the stability of degradation assessments. This method performs structured interval analysis on vibration signals and grades and encodes vibration impact intensity, making the degradation indices focus more on the overall trend of impact energy changes, rather than the random fluctuations of individual sampling points. Therefore, this method can maintain good stability in noisy environments and improve the reliability of degradation assessment results.

[0159] The degradation trend described in this invention is smoother and more stable, which is beneficial for condition monitoring and lifespan prediction: During equipment degradation, ideal health indicators should show a relatively stable trend over time. However, traditional vibration statistics often fluctuate frequently in noisy environments, resulting in unclear degradation trends. This method considers both vibration impact intensity and periodic structural characteristics when constructing degradation indicators, enabling the indicator changes to more accurately reflect the equipment degradation process, thus obtaining a smoother and more stable degradation curve. Therefore, the degradation indicators constructed by this method are more beneficial for equipment condition monitoring and remaining lifespan prediction.

[0160] The computational structure of this invention is simple and suitable for online monitoring scenarios of wind turbine units. This invention primarily constructs degradation indices based on vibration interval extreme value statistics, simple sorting operations, and weight calculations, without involving complex model training or large-scale numerical optimization. Therefore, this method has low computational complexity, is easy to deploy in real-time in wind turbine unit online monitoring systems, and can meet the needs of long-term operational status monitoring of wind turbine gearboxes.

[0161] Figure 4 A wind turbine gearbox degradation assessment system is shown, comprising:

[0162] The signal processing unit is configured to: acquire the original vibration signal of the wind turbine gearbox, preprocess the original vibration signal, divide the preprocessed vibration signal into multiple vibration intervals based on the zero crossover point, extract the extreme amplitude of each vibration interval, and obtain the interval extreme value sequence.

[0163] The amplitude encoding unit is configured to: construct an amplitude reference threshold based on historical vibration samples under healthy conditions, perform hierarchical mapping on the interval extreme value sequence according to the amplitude reference threshold, and generate an interval amplitude intensity encoding sequence;

[0164] The periodic quantization unit is configured to: perform autocorrelation analysis on the original vibration signal, identify the delay characteristics in the impact anomaly sequence based on the autocorrelation analysis results, and obtain the periodic intensity characteristics of the vibration interval;

[0165] The index calculation unit is configured to: synchronously sort the interval amplitude intensity coding sequence and periodic intensity features to obtain an ordered periodic intensity sequence; construct periodic weights based on the ordered periodic intensity sequence; and perform weighted correction on the Gini index based on the periodic weights to generate a periodic perception weighted Gini index.

[0166] The trend building unit is configured to repeatedly calculate the periodically perceived weighted Gini index based on a sliding time window to generate a degradation index sequence to characterize the degradation evolution process of wind turbine gearboxes.

[0167] It is understood that the aforementioned units can be individually or entirely merged into one or more other units, or some of the units can be further divided into multiple functionally smaller units. This achieves the same operation without affecting the technical effects of the embodiments of the present invention. The aforementioned units are based on logical functional division. In practical applications, the function of one unit can be implemented by multiple units, or the function of multiple units can be implemented by one unit. In other embodiments of the present invention, the system may also include other units. In practical applications, these functions can also be implemented with the assistance of other units, and can be implemented collaboratively by multiple units.

[0168] According to another embodiment of the present invention, the system of this embodiment can be constructed by running a computer program (including program code) capable of performing the steps involved in the corresponding method of the present invention on a general-purpose computing device, such as a computer, which includes processing elements and storage elements such as a central processing unit (CPU), random access memory (RAM), and read-only memory (ROM). The computer program can be recorded on, for example, a computer-readable recording medium, loaded into the aforementioned computing device through the computer-readable recording medium, and run therein.

[0169] Figure 5 A computer device is shown, comprising a processor, a communication interface, and a computer-readable storage medium. The processor, communication interface, and computer-readable storage medium are connected via a bus or other means.

[0170] The communication interface is used to receive and send data. The computer-readable storage medium can be stored in the memory of the electronic device. The computer-readable storage medium is used to store computer programs, which include program instructions. The processor is used to execute the program instructions stored in the computer-readable storage medium.

[0171] A processor is the computing and control core of an electronic device. It is suitable for implementing one or more instructions, specifically for loading and executing one or more instructions to achieve the corresponding method flow or function.

[0172] The processor is configured to perform the following procedure:

[0173] The original vibration signal of the wind turbine gearbox is acquired, the original vibration signal is preprocessed, and the preprocessed vibration signal is divided into multiple vibration intervals based on the zero crossover point. The extreme value amplitude of each vibration interval is extracted to obtain the interval extreme value sequence.

[0174] An amplitude reference threshold is constructed based on historical vibration samples under healthy conditions. The interval extreme value sequence is then hierarchically mapped according to the amplitude reference threshold to generate an interval amplitude intensity coding sequence.

[0175] Autocorrelation analysis is performed on the original vibration signal, and the delay characteristics in the impact anomaly sequence are identified based on the autocorrelation analysis results to obtain the periodic intensity characteristics of the vibration interval.

[0176] The interval amplitude intensity coding sequence and periodic intensity features are simultaneously sorted to obtain an ordered periodic intensity sequence. Periodic weights are constructed based on the ordered periodic intensity sequence, and the Gini index is weighted and corrected based on the periodic weights to generate a periodic perception weighted Gini index.

[0177] Based on the repeated calculation of the periodic perception-weighted Gini index using a sliding time window, a degradation index sequence is generated to characterize the degradation evolution process of wind turbine gearboxes.

[0178] This invention also provides a computer-readable storage medium, which is a memory device in an electronic device for storing programs and data. It is understood that the computer-readable storage medium here may include both built-in storage media in the electronic device and extended storage media supported by the electronic device. The computer-readable storage medium provides storage space for storing the processing system of the electronic device.

[0179] Furthermore, this storage space also contains one or more instructions suitable for loading and execution by the processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory; alternatively, it can also be at least one computer-readable storage medium located remotely from the aforementioned processor.

[0180] In one embodiment, the computer-readable storage medium stores one or more instructions; the processor loads and executes the one or more instructions stored in the computer-readable storage medium to perform the following process:

[0181] The original vibration signal of the wind turbine gearbox is acquired, the original vibration signal is preprocessed, and the preprocessed vibration signal is divided into multiple vibration intervals based on the zero crossover point. The extreme value amplitude of each vibration interval is extracted to obtain the interval extreme value sequence.

[0182] An amplitude reference threshold is constructed based on historical vibration samples under healthy conditions. The interval extreme value sequence is then hierarchically mapped according to the amplitude reference threshold to generate an interval amplitude intensity coding sequence.

[0183] Autocorrelation analysis is performed on the original vibration signal, and the delay characteristics in the impact anomaly sequence are identified based on the autocorrelation analysis results to obtain the periodic intensity characteristics of the vibration interval.

[0184] The interval amplitude intensity coding sequence and periodic intensity features are simultaneously sorted to obtain an ordered periodic intensity sequence. Periodic weights are constructed based on the ordered periodic intensity sequence, and the Gini index is weighted and corrected based on the periodic weights to generate a periodic perception weighted Gini index.

[0185] Based on the repeated calculation of the periodic perception-weighted Gini index using a sliding time window, a degradation index sequence is generated to characterize the degradation evolution process of wind turbine gearboxes.

[0186] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can implement the described functions using different methods for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0187] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of the present invention is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in or transmitted through a computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic cable, digital cable) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data processing device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive), etc.

[0188] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method of assessing the degradation level of a wind turbine gearbox, characterized in that, Includes the following processes: The original vibration signal of the wind turbine gearbox is acquired, the original vibration signal is preprocessed, and the preprocessed vibration signal is divided into multiple vibration intervals based on the zero crossover point. The extreme value amplitude of each vibration interval is extracted to obtain the interval extreme value sequence. An amplitude reference threshold is constructed based on historical vibration samples under healthy conditions. The interval extreme value sequence is then hierarchically mapped according to the amplitude reference threshold to generate an interval amplitude intensity encoding sequence, including: Calculate the high quantiles of the extreme value sequences of multiple historical health status samples, and fuse the high quantiles to obtain the health reference threshold; Each extreme value in the interval extreme value sequence is compared with a health reference threshold. Based on the comparison results, the extreme values ​​are mapped to different intensity levels to generate an interval amplitude intensity coding sequence. Anomaly thresholds are constructed based on health reference thresholds and the absolute median difference method. Based on these thresholds, interval amplitude intensity encoded sequences are detected to generate impact anomaly sequences. Autocorrelation analysis is performed on the original vibration signals, and delay features in the impact anomaly sequences are identified based on the autocorrelation analysis results, yielding the periodic intensity characteristics of the vibration intervals, including: Calculate the autocorrelation function of the shock anomaly sequence, and estimate the shock period based on the location of the maximum value of the autocorrelation function; Based on the impact period, the consistency of the impact anomaly sequence is calculated in the neighborhood, and the periodic intensity characteristics of the vibration interval are obtained. The interval amplitude intensity coding sequence and periodic intensity features are simultaneously sorted to obtain an ordered periodic intensity sequence. Periodic weights are constructed based on the ordered periodic intensity sequence, and the Gini index is weighted and corrected based on the periodic weights to generate a periodic perception weighted Gini index. Based on the repeated calculation of the periodic perception-weighted Gini index using a sliding time window, a degradation index sequence is generated to characterize the degradation evolution process of wind turbine gearboxes.

2. The method for assessing the degradation degree of wind turbine gearboxes as described in claim 1, characterized in that, Based on the comparison results, extreme values ​​are mapped to different intensity levels, generating interval amplitude intensity coding sequences, including: The extreme value of the interval is compared with the health reference threshold. If the extreme value of the interval is less than or equal to the health reference threshold, the first intensity level code is assigned. If the extreme value of the interval is greater than the health reference threshold, then calculate the multiple by which the extreme value of the interval exceeds the health reference threshold; Based on the preset multiple interval division rules, different multiple intervals are mapped to different intensity level codes, generating interval amplitude intensity code sequences.

3. The method for assessing the degradation degree of wind turbine gearboxes as described in claim 1, characterized in that, The interval amplitude intensity coding sequence and periodic intensity features are simultaneously sorted to obtain an ordered periodic intensity sequence. Periodic weights are constructed based on this ordered periodic intensity sequence, and the Gini index is weighted and corrected based on these periodic weights to generate a periodically-aware weighted Gini index, including: The interval amplitude intensity coding sequence is sorted in ascending order to obtain an ordered amplitude sequence. Based on the mapping relationship between the interval amplitude intensity coding sequence and the ordered amplitude sequence, the periodic intensity features are rearranged to obtain an ordered periodic intensity sequence. The weight coefficients for each interval are calculated based on the ordered periodic intensity sequence to obtain the periodic weight sequence; Based on the periodic weight sequence and the ordered amplitude sequence, the periodic perceived weighted Gini index is calculated through weighted summation and accumulation operations.

4. The method for assessing the degradation degree of wind turbine gearboxes as described in claim 3, characterized in that, Based on the periodic weight sequence and ordered magnitude sequence, the periodic-perceived weighted Gini exponent is calculated through weighted summation and accumulation operations, including: To calculate the weighted cumulative sum, multiply each element in the ordered magnitude sequence by its corresponding period weight, and accumulate all the product results from the first element to the current element. Calculate the weighted sum of the total energy, and sum the products of all ordered magnitudes and their corresponding period weights; Construct the weighted area under the Lorentz curve based on the ratio of the weighted cumulative sum to the weighted sum of total energy. The periodic perceived weighted Gini index is obtained by calculating the non-uniformity of the numerical distribution based on the weighted area.

5. The method for assessing the degradation degree of wind turbine gearboxes as described in claim 1, characterized in that, In the step of generating an interval amplitude intensity encoded sequence by hierarchically mapping the interval extreme value sequence according to the amplitude reference threshold, the logic of hierarchical mapping is as follows: When the extreme value of the interval is less than or equal to the health reference threshold, the output code is 0; When the extreme value of the interval is greater than the health reference threshold but less than or equal to twice the health reference threshold, output code 1; When the extreme value of the interval is greater than twice the health reference threshold and less than or equal to three times the health reference threshold, output code 2; When the extreme value of the interval is greater than 3 times the health reference threshold and less than or equal to 4 times the health reference threshold, output code 3; When the extreme value of the interval is greater than 4 times the health reference threshold, the output code is 4; Arrange the above codes in chronological order to generate an interval amplitude intensity coding sequence.

6. A wind turbine gearbox degradation assessment system employing the wind turbine gearbox degradation assessment method as described in any one of claims 1-5, characterized in that, include: The signal processing unit is configured to: acquire the original vibration signal of the wind turbine gearbox, preprocess the original vibration signal, divide the preprocessed vibration signal into multiple vibration intervals based on the zero crossover point, extract the extreme amplitude of each vibration interval, and obtain the interval extreme value sequence. The amplitude encoding unit is configured to: construct an amplitude reference threshold based on historical vibration samples under healthy conditions, perform hierarchical mapping on the interval extreme value sequence according to the amplitude reference threshold, and generate an interval amplitude intensity encoding sequence; The periodic quantization unit is configured to: perform autocorrelation analysis on the original vibration signal, identify the delay characteristics in the impact anomaly sequence based on the autocorrelation analysis results, and obtain the periodic intensity characteristics of the vibration interval; The index calculation unit is configured to: synchronously sort the interval amplitude intensity coding sequence and periodic intensity features to obtain an ordered periodic intensity sequence; construct periodic weights based on the ordered periodic intensity sequence; and perform weighted correction on the Gini index based on the periodic weights to generate a periodic perception weighted Gini index. The trend building unit is configured to repeatedly calculate the periodically perceived weighted Gini index based on a sliding time window to generate a degradation index sequence to characterize the degradation evolution process of wind turbine gearboxes.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program adapted to be loaded by a processor and executed as described in any one of claims 1 to 5 for assessing the degree of degradation of a wind turbine gearbox.

8. A computer device, characterized in that, include: Processor and computer-readable storage media; A processor, adapted to execute computer programs; A computer-readable storage medium storing a computer program that, when executed by the processor, implements the wind turbine gearbox degradation assessment method as described in any one of claims 1 to 5.

Images