A method and system for constructing insulation degradation characteristics of generator motors

Abstract

This invention relates to the field of generator motor technology, and particularly to a method and system for constructing insulation degradation characteristics of generator motors. The method acquires three-phase stator current signals and operating mode identification signals from the generator motor. After signal preprocessing, it performs sequence component decomposition and frequency domain analysis to extract zero-sequence components, negative-sequence components, and specific order harmonic amplitudes, constructing a feature set related to the insulation degradation mechanism. Addressing the statistical differences in feature amplitudes between generator and motor modes, the method calculates the mean value of features under both operating conditions and constructs a mode sensitivity factor, establishing a proportional mapping function to align feature scales. Based on this, a mechanism weighting coefficient is introduced to construct a structured insulation degradation feature matrix. This method can eliminate the influence of operating mode differences on feature expression, improve the consistency and stability of insulation degradation characteristics under dual operating conditions, and is suitable for online condition assessment and trend analysis of pumped storage generator motors.

Description

Technical Field

[0001] This invention relates to the field of generator motor technology, and in particular to a method and system for constructing insulation degradation characteristics of generator motors. Background Technology

[0002] With the high proportion of new energy grid connection and the continuous improvement of energy electrification, the power system is exhibiting stronger volatility and uncertainty under the guidance of the "dual carbon" target. Pumped storage units, with their large capacity and fast start-up and shutdown characteristics, have become an important support for power grid peak shaving and valley filling, primary / secondary frequency regulation, and inertia support. Pumped storage units frequently switch between power generation and motoring modes, resulting in significant multi-field coupling effects of hydraulic, electromagnetic, and mechanical fields. The operating signals exhibit complex characteristics of non-stationarity, strong noise, and superposition across time scales, which significantly increases the technical difficulty of condition assessment and early anomaly detection.

[0003] As the core equipment of pumped storage units, the generator motor is a reversible electromechanical device that combines power generation mode and electric mode. It can switch between the two operating states to achieve bidirectional conversion of electrical energy and mechanical energy. The insulation system of the generator motor is a component that ensures the safe and reliable operation of the motor. It mainly includes: inter-strand insulation, inter-turn insulation, inter-layer insulation, inter-phase insulation, and insulation to ground. The insulation system of the generator motor is prone to gradual deterioration under the coupling effects of long-term electromagnetic stress, thermal gradient changes, and mechanical shock.

[0004] Current research on the insulation status of pumped storage units mainly includes the following aspects:

[0005] (1) Directly collect signals under the operating state of pumped storage units to construct health status characteristics, such as the Chinese invention patent with application number CN202210156063.0 and title "Method and device for creating health status model and predicting performance trend of pumped storage units". This patent proposes a trend prediction idea based on the sequence of health performance indicators, but its technical path focuses on constructing a prediction model based on preset indicators.

[0006] (2) Consider the insulation status of the stator of the pumped storage unit, such as the patent application with application number CN202511464793.7 entitled "Method for monitoring the insulation status of stator winding of unit based on neutral point harmonic sweep frequency injection" and the patent application with application number CN202511465380.0 entitled "Online monitoring method for the insulation status of stator winding of full power variable speed pumped storage unit".

[0007] (3) Consider the insulation status of the rotor of the pumped storage unit, such as the patent application with application number CN202511037320.9 entitled "An insulation status monitoring system and method for the rotor winding of a pumped storage unit", and the patent application with application number CN202410618270.2 entitled "An insulation status monitoring system and method for the rotor winding of a variable speed pumped storage unit".

[0008] However, the above research has the following problems: it lacks a systematic feature construction process oriented towards mechanism constraints for the harmonic and sequence component information in the original electrical signal that is highly related to the insulation degradation mechanism. At the same time, it has not established a consistent mapping or alignment mechanism for the feature distribution differences caused by frequent switching between power generation and motoring conditions, resulting in inaccurate degradation feature construction. This is because, from the perspective of motor fault mechanism analysis, insulation degradation often destroys the structural symmetry of the electromagnetic system, thereby inducing amplitude and phase changes in the negative sequence component, zero sequence component, and specific order harmonic components in the three-phase current.

[0009] Therefore, mining harmonic and sequence component characteristics based on raw electrical quantities such as three-phase current is an important technical approach for characterizing insulation degradation mechanisms. However, under conditions of frequent switching between power generation and motoring, the statistical distribution and amplitude sensitivity of the same type of harmonic or imbalance index vary significantly under different operating modes. Without a distribution alignment or consistency mapping mechanism for operating mode switching, the characteristics are easily masked by transient disturbances or random fluctuations during switching, resulting in a discrete distribution of the characteristic space across different operating conditions, making it difficult to form a unified and stable degradation characterization.

[0010] In current engineering practice, insulation condition assessment often relies on the time-domain / frequency-domain statistical characteristics of single or limited signals such as current, temperature, and vibration, combined with threshold criteria or shallow models for judgment. While this approach is applicable under stable operating conditions or low noise, in scenarios with frequent switching between multiple operating conditions, rapid load changes, and alternating start-stop / switching conditions, issues such as signal mode aliasing, feature redundancy, and statistical distribution drift can easily lead to unstable feature representation. Threshold criteria can also shift with fluctuations in operating conditions, making it difficult to stably characterize the degradation process across operating conditions and over long time periods, thus weakening the ability to identify latent early insulation degradation.

[0011] Therefore, from both theoretical research and engineering practice perspectives, it is still necessary to propose a feature construction method for generator motor insulation degradation, which is based on mechanism-constrained harmonic characteristics and has dual-condition consistency mapping capability, in order to meet the engineering application requirements for stable and interpretable feature expression under complex operating conditions of pumped storage units. Summary of the Invention

[0012] Technical Problem: To address the technical problems existing in the prior art, this invention provides a method for constructing insulation degradation characteristics of a generator motor. This method solves the problem that, under the frequent switching conditions of pumped storage units operating in dual modes, the statistical distribution differences and amplitude sensitivity changes of electrical characteristics such as harmonics and sequence components in different operating modes make it difficult to establish a unified and stable insulation degradation characterization space. Furthermore, it solves the problem that in non-stationary, high-noise, and multi-scale coupled operating environments, insulation degradation characteristics are prone to drift and inconsistent expression, thus affecting the reliability and generalization ability of subsequent insulation condition assessment and trend analysis. In addition, this application also provides a system for constructing insulation degradation characteristics of a generator motor.

[0013] Technical Solution: According to the technical solution provided by this invention, on one hand, this invention provides a method for constructing insulation degradation characteristics of a generator motor, the method comprising:

[0014] During the stable operation of the pumped storage unit, the three-phase voltage signal and three-phase stator current signal of the pumped storage unit are collected synchronously, and the operating status of the pumped storage unit is collected synchronously as an operating mode identification signal. The operating mode identification signal is used to indicate whether the pumped storage unit is in power generation mode or electric mode.

[0015] After preprocessing the acquired three-phase stator current signal, a preprocessed windowed three-phase stator current sequence is obtained. The zero-sequence component, positive-sequence component, and negative-sequence component of the preprocessed windowed three-phase stator current sequence in each time window are extracted, and the harmonic features of several orders of the preprocessed windowed three-phase stator current sequence in each time window are extracted to construct an insulation degradation feature vector.

[0016] Based on the insulation degradation feature vector, a mode sensitivity factor is determined. The mode sensitivity factor is then combined with a consistency mapping function to perform scale unification on the insulation degradation feature vector to obtain a cross-mode consistent insulation degradation feature sequence. The mode sensitivity factor is used to characterize the scale ratio of the electric mode to the power generation mode corresponding to each dimension feature in the insulation degradation feature vector.

[0017] Based on the cross-mode consistent feature sequence, a structured insulation degradation feature matrix is ​​constructed, and the final feature matrix is ​​obtained by using a mechanism constraint method, thereby performing insulation status assessment and early warning output.

[0018] Furthermore, including:

[0019] The step of extracting harmonic features of several orders of the preprocessed windowed three-phase stator current sequence in each time window and then constructing an insulation degradation feature vector includes: using Fast Fourier Transform to extract the amplitude of several orders of harmonic features of the preprocessed windowed three-phase stator current sequence in each time window, and determining the target harmonic order set based on the set mechanism prior constraint method; and concatenating the amplitudes of the negative sequence component, the zero sequence component, and the harmonic features of the determined order to construct the insulation degradation feature vector.

[0020] Furthermore, including:

[0021] The method for determining the target harmonic order set based on the set mechanistic prior constraints includes:

[0022] A candidate order set is constructed based on the symmetry destruction and dielectric nonlinear enhancement mechanism caused by insulation degradation. The symmetry destruction corresponds to the enhancement of negative or zero-sequence components and is accompanied by changes in even-order harmonic components. The dielectric nonlinear enhancement corresponds to changes in several odd-order harmonic components.

[0023] Based on the statistical screening of each candidate order in the candidate order set using sample windows, the harmonic amplitude sequence corresponding to the sample window is divided into a healthy segment and a suspected degradation segment. The discrimination index and trend correlation index are used to determine the discrimination and degradation trend correlation of any candidate order in the healthy segment and the suspected degradation segment, respectively. The healthy segment is selected from the window set where the insulation status is confirmed to be normal after maintenance or in the early stage of operation, or the window set where the negative sequence component, zero sequence component and harmonic level are within the baseline threshold range and there are no abnormal alarms. The suspected degradation segment is selected from the window set where there are offline detection or maintenance records indicating degradation, or the window set that meets the degradation criteria.

[0024] The current comprehensive score is determined based on the correlation between discrimination and degradation trend, and candidate orders with scores below the first threshold are eliminated in order to retain orders that are more strongly associated with degradation mechanism.

[0025] To eliminate orders that are highly sensitive to load fluctuations but weakly correlated with degradation trends, the correlation between the amplitude of the harmonic characteristics of the current order and the load characterization quantity is calculated and denoted as characterization correlation. If the characterization correlation is higher than the second threshold and the comprehensive score of the corresponding order is lower than the first threshold, then the order is removed from the candidate order set.

[0026] To ensure the stability of the selected harmonic order under different operating conditions, the comprehensive score and the characterization correlation are repeatedly calculated in power generation mode, electric mode, and different load ranges to obtain the final target harmonic order set.

[0027] Furthermore, including:

[0028] The determination of the mode sensitivity factor based on the insulation degradation feature vector includes:

[0029] Determine the dimension of the insulation degradation feature vector Its dimension is obtained by adding the zero-sequence component, the negative-sequence component, and the order of the target harmonic;

[0030] Based on the operating mode identification signal, all sliding time windows are divided into a power generation mode window set and a motoring mode window set. The statistical scaling parameter for each feature dimension is calculated for both the power generation mode window set and the motoring mode window set, and is expressed as follows: ;

[0031] in, Indicates the first Each dimension corresponds to the mean scale of the feature under the power generation mode. Indicates the first The mean scale of the dimensional feature in the electric mode. These represent the number of windows within the power generation mode window set and the electric mode window set, respectively. Indicates the sliding time window number. For the first Feature vectors corresponding to each dimension;

[0032] Definition of the first The pattern sensitivity factor corresponding to each dimension of the feature is represented as: ;

[0033] in, Indicates the first The scale ratio of the features corresponding to each dimension in electric mode relative to power generation mode; It is a tiny positive number that approaches zero, used to prevent the denominator from being zero.

[0034] Furthermore, including:

[0035] The step of using the mode sensitivity factor combined with the consistency mapping function to scale the insulation degradation feature vector to obtain a cross-mode consistent insulation degradation feature sequence includes:

[0036] Using a consistent mapping function to represent the first The feature corresponding to the dimension in the th dimension is at the th . The values ​​within each sliding time window are: Thus, a cross-mode consistent insulation degradation characteristic sequence is obtained. The consistency mapping is used to mitigate the systematic scale shift caused by the difference between the power generation mode and the electric mode. It only adjusts the amplitude scale of each feature dimension to achieve scale alignment.

[0037] Furthermore, including:

[0038] Based on the cross-mode consistent feature sequences, a structured insulation degradation feature matrix is ​​constructed, including:

[0039] Verify the alignment effect of the consistency mapping on the scale shift caused by the difference between power generation and electric modes, i.e., for any feature dimension. ,Depend on And when hour, get: ;

[0040] Therefore, after the consistency mapping, the mean scale of the same feature dimension in the two operating modes becomes consistent in a statistical sense;

[0041] Let the total number of time windows used for analysis be . ,in, The window numbers are arranged in chronological order, and the mapped feature vectors for each window are... Construct an insulation degradation feature matrix by splicing the data in chronological order. ,for: ;

[0042] in, The rows correspond to different feature dimensions, and the columns correspond to sliding time window samples arranged in chronological order.

[0043] Furthermore, including:

[0044] The mechanism constraint method includes:

[0045] Set a mechanism priority coefficient for each dimension's corresponding feature. Among them, the priority coefficients corresponding to the negative sequence component characteristics and the zero sequence component characteristics are higher than the priority coefficients corresponding to the amplitude characteristics of the target harmonic.

[0046] Set stability suppression coefficient To reduce the weight of features that are more sensitive to fluctuations in the running point, the stability suppression coefficient is determined by the degree of dispersion of the feature;

[0047] The sensitivity weights of each dimension are obtained by multiplying the priority coefficient and the stability inhibition coefficient by the computer and normalizing the product.

[0048] Dimensions based on insulation degradation feature vectors Define the mechanism weight vector as follows: ,in, Indicates the first Each dimension corresponds to a feature's sensitivity weight to the insulation degradation mechanism; the larger the weight, the stronger the correlation between the feature and the insulation degradation mechanism.

[0049] Based on the aforementioned mechanism weight vector, the weighted final feature matrix is ​​constructed as follows: ;in, Represents the mechanism weight vector A diagonal matrix constructed for the diagonal elements.

[0050] Furthermore, including:

[0051] The stability suppression coefficient is expressed as: ;in, For the first Each dimension corresponds to the coefficient of variation of the feature over the entire sliding time window. and These are the mean and standard deviation of this feature dimension, respectively. It is a tiny positive number that approaches zero.

[0052] According to the technical solution provided by the present invention, in a second aspect, the present invention provides a system for constructing insulation degradation characteristics of a generator motor, the system comprising:

[0053] The data acquisition module is used to simultaneously acquire the three-phase voltage signal and three-phase stator current signal of the pumped storage unit during stable operation, and to simultaneously acquire the operating status quantity of the pumped storage unit as an operating mode identification signal. The operating mode identification signal is used to indicate whether the pumped storage unit is in power generation mode or electric mode.

[0054] An insulation degradation feature vector construction module is used to preprocess the acquired three-phase stator current signal to obtain a preprocessed windowed three-phase stator current sequence, extract the zero-sequence component, positive-sequence component and negative-sequence component of the preprocessed windowed three-phase stator current sequence in each time window, and extract the harmonic features of the preprocessed windowed three-phase stator current sequence in each time window to construct an insulation degradation feature vector.

[0055] An insulation degradation feature sequence construction module is used to determine a mode sensitivity factor based on the insulation degradation feature vector, and use the mode sensitivity factor in combination with a consistency mapping function to perform scale unification on the insulation degradation feature vector to obtain a cross-mode consistent insulation degradation feature sequence. The mode sensitivity factor is used to characterize the scale ratio of the electric mode to the power generation mode corresponding to each dimension feature in the insulation degradation feature vector.

[0056] The feature matrix determination module is used to construct a structured insulation degradation feature matrix based on the cross-mode consistent feature sequence, and to obtain the final feature matrix using a mechanism constraint method, thereby performing insulation status assessment and early warning output.

[0057] According to the technical solution provided by the present invention, in a third aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the method described in any of the above-mentioned embodiments.

[0058] Beneficial Effects: This invention proposes a method and system for constructing insulation degradation features of generator-motor units based on mechanistic constraint harmonic feature extraction and dual-condition consistency mapping fusion. Compared to methods that construct features only under a single operating condition or directly mix features from both operating conditions, this invention constructs a mode sensitivity factor based on the collected operating mode identifiers. This factor is used to represent the scale ratio of the electric mode to the generator mode corresponding to each dimension of the insulation degradation feature vector. Based on the mode sensitivity factor, a consistency mapping relationship between the generator mode and the electric mode features is established, thereby achieving scale alignment and unified expression of the sequence component features and the target harmonic features under both operating conditions. Furthermore, this invention introduces a mechanism weight coefficient to reconstruct key features, strengthening feature components highly correlated with the insulation degradation mechanism and suppressing weakly correlated features more sensitive to fluctuations in operating points, forming a structured insulation degradation feature matrix. This improves the stability, consistency, and interpretability of feature expression under alternating generator and electric operation conditions and complex operating conditions of pumped storage units, providing more reliable feature input for subsequent insulation condition assessment, anomaly detection and early warning, and trend analysis. Attached Figure Description

[0059] Figure 1 This is a flowchart of the method for constructing insulation degradation characteristics of a generator motor according to an embodiment of the present invention;

[0060] Figure 2 This is a schematic diagram illustrating the feature mapping process before and after using the consistency mapping function as described in an embodiment of the present invention.

[0061] Figure 3 This is a schematic diagram of the system structure for constructing insulation degradation characteristics of a generator motor according to an embodiment of the present invention. Detailed Implementation

[0062] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0063] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0064] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0065] Please see Figure 1 This invention provides a method for constructing insulation degradation features of a generator-motor. This method uses sequence component decomposition and harmonic order extraction to structure electrical quantity signals such as three-phase current into a set of sequence component-harmonic features related to the insulation degradation mechanism. It then combines operating mode information to construct a mode-sensitive factor, establishing a feature consistency mapping relationship across generation and motoring modes. This projects harmonic features under different operating conditions onto a unified feature space, forming a feature vector or feature matrix consistent across both operating conditions and with mechanism constraints, thus achieving stable construction and interpretable expression of insulation degradation features. The method includes the following steps:

[0066] During the stable operation of the pumped storage unit, S1 synchronously collects the three-phase voltage signal and three-phase stator current signal on the stator side of the pumped storage unit, and synchronously collects the operating status quantity of the pumped storage unit as the operating mode identification signal. The operating mode identification signal is used to indicate whether the pumped storage unit is in power generation mode or electric mode.

[0067] In a preferred embodiment, during the stable operation of the pumped storage unit, the three-phase voltage signal on the stator side of the unit is collected synchronously. With three-phase stator current signal Simultaneously collect unit operating status data as operating mode identification signals. The operating mode identification signal is mentioned above. It can be obtained from the operating status bit, power direction indication or active power measurement output by the unit monitoring system or protection / control device, and is used to indicate whether the unit is in power generation mode or electric mode, i.e. pumping mode. It is also timestamped with the above electrical quantity signals to form a synchronous data stream of "electrical quantity-operating status tag".

[0068] Preferably, the three-phase current signal is acquired through a current transformer installed at the stator winding outlet, with a sampling frequency of [missing information]. The frequency range is selected from 1kHz to 10kHz to cover the target harmonic frequency band and meet the requirements of subsequent frequency domain analysis. The sampling data adopts a unified clock or time synchronization mechanism, such as synchronous sampling on the same acquisition card, PTP / IRIG-B time synchronization, or software resampling alignment, to ensure the phase consistency and comparability of the three-phase electrical quantities.

[0069] Furthermore, the raw signals acquired synchronously are processed using a sliding window slicing method: Let the window length be... Number of sampling points, sliding step size is Each sampling point divides the continuous signal into a sequence window. , where each window To ensure the stability of subsequent pattern discrimination and graph structure modeling, the following preprocessing and quality control are performed within each window:

[0070] (1) DC component removal and amplitude standardization: The average value of each phase voltage and current is removed to eliminate measurement bias, and normalized according to the rated value or window effective value to make the dimensions consistent under different operating conditions / different acquisition channels.

[0071] (2) Abnormal window removal: When abnormalities such as sampling loss, saturation truncation, and sudden pulses occur in the window, such as exceeding the mean or exceeding the rated value threshold, the window is marked as invalid and removed or reduced in weight.

[0072] (3) Sampling Synchronization Correction: The sampling delay in the three-phase channels is corrected to ensure that all phase quantities within the same window correspond at the same sampling time. The preprocessed windowed data is obtained through the above steps. .

[0073] S2 preprocesses the acquired three-phase stator current signal to obtain a preprocessed windowed three-phase stator current sequence. It then extracts the zero-sequence component, positive-sequence component, and negative-sequence component of the preprocessed windowed three-phase stator current sequence in each time window, as well as extracts several harmonic features of the preprocessed windowed three-phase stator current sequence in each time window, and then constructs an insulation degradation feature vector.

[0074] Among them, the obtained preprocessed windowed three-phase stator current sequence In each time window The inner sequence components and harmonic characteristics are extracted to characterize the electrical asymmetry and nonlinear enhancement effects caused by insulation degradation. Preferably, in the first... Within each time window, the three-phase current sequence at the fundamental frequency was analyzed. Discrete Fourier analysis was performed at the point to obtain the fundamental phasor of the three-phase current. The fundamental phasor can be obtained by FFT / DFT at a frequency of The complex coefficients at the given location are obtained.

[0075] Furthermore, a symmetric component transformation is performed on the fundamental phasor to obtain the zero-sequence component, the positive-sequence component, and the negative-sequence component: (1)

[0076] in, For zero-order components, For positive order components, For negative order components, This is the phase rotation constant. Insulation degradation, such as inter-turn insulation deterioration, can cause winding impedance asymmetry and disrupt three-phase symmetry, thereby increasing the unbalance component. Therefore, the negative sequence component is preferred. and zero-order components As a degradation-sensitive characteristic; in contrast, the positive-order component It mainly reflects the fundamental main component under normal symmetrical operation, and is more susceptible to load, excitation and operating point fluctuations. It also contributes less to the characterization of asymmetric enhancement. In order to highlight the symmetry destruction effect caused by degradation, this embodiment does not take the positive sequence component as the core characteristic quantity.

[0077] In terms of harmonic feature extraction, in the first Within a time window, perform a Fast Fourier Transform on the preprocessed discrete current sequence to extract the first... First harmonic amplitude Its expression is: (2)

[0078] in, Preferably, any phase current sequence can be used; more preferably, a three-phase composite current or a sequence component current sequence can also be used to enhance robustness. The fundamental frequency, The order of the target harmonic.

[0079] Regarding the determination of the target harmonic order, to ensure that the selected harmonic components have a stable correlation with the insulation degradation mechanism, a target harmonic order set is required. The preferred approach is to determine the order by combining "mechanism prior constraints with statistical screening and verification": firstly, a set of candidate orders is given based on the mechanism of symmetry failure and nonlinear enhancement caused by insulation degradation. Then, based on historical operating data or sample windows, the discrimination or trend correlation index of candidate harmonic components between healthy and suspected degradation segments is calculated, and orders that are highly sensitive to load fluctuations but have a weak correlation with degradation trends are eliminated. Finally, the stability of the selected orders is tested in both power generation and electric operation modes and in different load ranges to obtain the final target order set. ,in .

[0080] Specifically, in a preferred embodiment of this invention, regarding the determination of the target harmonic order, to ensure that the selected harmonic components have a stable correlation with the insulation degradation mechanism and to reduce the interference of load fluctuations on the order selection, a target harmonic order set is established. The preferred approach is to combine mechanistic prior constraints with statistical screening and verification. The specific steps are as follows:

[0081] First, a candidate order set is constructed based on the symmetry destruction and dielectric nonlinear enhancement mechanism caused by insulation degradation. Among these, symmetry violation typically corresponds to enhancement of negative / zero-sequence components accompanied by changes in even-order harmonic components, while nonlinear enhancement typically corresponds to changes in several odd-order harmonic components; therefore, Preferred from Selected from, among which, The order is determined based on the sampling frequency and effective frequency band, and can preferentially include lower-order harmonics such as the 2nd, 3rd, and 4th harmonics based on experience with motor structure and mechanism, but the order is not limited to these, thus forming a candidate order set. .

[0082] Then, statistical filtering is performed on each candidate order based on historical operating data or a sample window. For this purpose, the sample window is divided into healthy segments. With suspected degeneration segment The healthy segment A set of windows can be selected after maintenance and during the initial operation period when the insulation condition is confirmed to be normal, or a set of windows can be selected where the negative sequence component, zero sequence component, and harmonic levels are within the baseline threshold range and there are no abnormal alarms; the suspected degradation segment A set of windows representing degradation based on offline detection and maintenance records can be selected, or a set of windows satisfying degradation criteria such as a continuous increase in negative or zero-order components and a continuous decrease in the insulation health index can be selected. For any candidate order... ,exist and The harmonic amplitude sequences are calculated respectively. The discriminant strength is correlated with the degradation trend, and the standardized difference between the healthy segment and the suspected degradation segment can be preferentially used as the discriminant index:

[0083] (3)

[0084] in, Within the healthy segment The mean and standard deviation, For the corresponding statistics within the suspected degradation segment, It is a very small positive number; the trend correlation index can preferably be the reference value for the degree of degradation. The absolute value of the Spearman correlation coefficient is expressed as:

[0085] (4)

[0086] in, This can be a degradation level mapped to offline detection metrics or maintenance records; when a degradation label is missing, it can be... Replace with an insulation health index or time index to characterize the direction of degradation evolution.

[0087] A comprehensive screening score is constructed based on the above indicators. ,in The weighting coefficients are used, and scores are preferentially excluded. Candidate orders below a threshold are selected to retain those more strongly associated with the degradation mechanism. This threshold is used to screen candidate orders that are highly associated with the insulation degradation mechanism and can be set empirically.

[0088] Furthermore, to eliminate orders that are highly sensitive to load fluctuations but weakly correlated with degradation trends, the amplitude of candidate orders is calculated. Compared with load characteristics, such as active power or fundamental amplitude Correlation between them:

[0089] (5)

[0090] when If the score exceeds a threshold, which is used to identify candidate orders that are overly sensitive to load fluctuations, this threshold can be set empirically and is determined by the overall score. When the threshold is low, that is, if the characterization correlation is higher than the second threshold and the comprehensive score of the corresponding order is lower than the first threshold, it is determined that the order mainly reflects load fluctuation rather than degradation trend, and thus it is removed from the candidate set to avoid mistakenly selecting the load-sensitive order as the target order.

[0091] Finally, to ensure the stability of the selected order under different operating conditions, the calculations were repeated in both power generation and electric operation modes, as well as in different load ranges. and duplicate calculations In this embodiment, different load ranges can be classified according to the absolute value of active power. The partitioning only retains the order where the threshold is met in each operating mode and each load partition, and the statistical fluctuation does not exceed the threshold. Statistical fluctuation not exceeding the threshold includes, for example, variance or coefficient of variation not exceeding the threshold.

[0092] (6)

[0093] Therefore, the final target harmonic order set is obtained.

[0094] Based on the above-mentioned sequence component characteristics and target harmonic amplitude characteristics, an insulation degradation feature vector is constructed:

[0095] (7)

[0096] in, It is used to characterize the unbalance enhancement and harmonic anomalies caused by insulation degradation, and serves as input for subsequent construction mode sensitivity factors and consistency mapping.

[0097] S3 determines the mode sensitivity factor based on the insulation degradation feature vector, and uses the mode sensitivity factor in combination with the consistency mapping function to perform scale unification on the insulation degradation feature vector to obtain a cross-mode consistent insulation degradation feature sequence. The mode sensitivity factor is the scale ratio of the electric mode to the power generation mode corresponding to each dimension feature in the insulation degradation feature vector.

[0098] In this embodiment, due to the differences in energy flow, magnetic field distribution, and operating point between the power generation mode and the electric mode, the electrical characteristic amplitudes corresponding to the same degree of insulation degradation may exhibit systematic scale shifts. To reduce the interference of operating mode differences on subsequent degradation identification and modeling, this embodiment uses a constructed insulation degradation feature vector. The pattern sensitivity factor is calculated, and the features are scaled uniformly through consistency mapping.

[0099] Insulation degradation feature vector The feature vector is ,in, Indicates the sliding time window number. The feature dimension; preferably, the feature dimension It equals the sum of the number of sequence component characteristics and the number of target harmonic characteristics, for example, when using and Target harmonic amplitude Sometimes, .therefore, The first eigenvector used to represent the eigenvector One dimension, It is not limited to 3 dimensions; the specific dimension varies depending on the number of harmonic orders of the selected target.

[0100] Furthermore, based on the operating mode identification information in step S1, all sliding time windows are divided into a power generation mode window set. Electric mode window set ,in, The set of window numbers that satisfy the condition "the window is in power generation mode". The set of window numbers that satisfy the condition "window is in electric mode" These represent the number of windows within the corresponding sets. To characterize the scale difference of the same feature dimension under the two operating modes, the statistical scale parameter of each feature dimension is calculated on both the power generation mode window set and the electric power mode window set. The mean is preferably used as the scale representation, resulting in:

[0101] (8)

[0102] in, Indicates the first The mean scale of the dimensional feature under the power generation mode. Indicates the first The mean scale of the feature in the electric mode. It should be noted that the summation in formula (8) is applied to the set of window indices. or Statistical analysis was performed on the sample window in the middle. The window number is used only to index the sample window, not to represent consecutive physical moments, thus avoiding ambiguity caused by mixing time with the sample set.

[0103] Based on the above mean scale, the first... The pattern sensitivity factors corresponding to the features in each dimension are:

[0104] (9)

[0105] in, Indicates the first The scale ratio of the dimensional feature in the electric mode relative to the power generation mode; Small positive numbers are used to avoid Too small a value will result in a denominator of zero or unstable values. The sensitivity factor is defined as the ratio of "electricity / generation" so that the generation mode can be used as a reference scale in the subsequent consistency mapping, and the characteristic amplitude of the electric mode can be scaled to a scale consistent with the generation mode. If the electric mode is used as the reference scale, the reciprocal of the sensitivity factor can be taken and the mapping direction can be adjusted accordingly. This embodiment does not limit its essential idea by the direction of the ratio. The key is to select a reference scale and keep the mapping rules consistent.

[0106] Furthermore, to achieve scale uniformity across operating modes, a consistent mapping of feature vectors is constructed. And a consistent mapping of features is performed using a dimension-wise scaling method, specifically:

[0107] (10)

[0108] in, Indicates the first consistent mapping 3D features in the 1st dimension The values ​​taken within a sliding time window, according to formula (10), when the window... Belongs to the power generation mode window set When the window is in a certain state, its characteristics remain unchanged; when the window belongs to the set of motorized mode windows... When using pattern-sensitive factors The first of the windows The dimensional features are scaled to map their scale to the reference scale of the power generation mode. This consistent mapping mitigates the systematic scale shift caused by differences between power generation and electric modes, improving the stability of subsequent degradation identification and modeling. Ultimately, a cross-mode consistent insulation degradation feature sequence is obtained. .

[0109] Based on the cross-mode consistent feature sequence, S4 constructs a structured insulation degradation feature matrix and uses a mechanism constraint method to obtain the final feature matrix, thereby performing insulation status assessment and early warning output.

[0110] Due to the electric mode window set Features within are mapped using a dimension-wise scaling method. and the power generation mode window set The internal features remain unchanged Therefore, the "scale correction" of the consistency mapping is specifically reflected in the following: only the amplitude scale of each feature in the electric mode is adjusted so that it is mapped to the reference scale of the power generation mode, thereby weakening the systematic scale shift caused by the difference in the operating mode; the mapping does not change the time window order, does not mix the calculation of different feature dimensions, and therefore does not introduce the coupling relationship between features, only corrects the amplitude scale and keeps the original time evolution structure unchanged.

[0111] Preferably, the scaling effect can be verified by the consistency of the mean scale of the mapped features across the two operating mode window sets. For any feature dimension... The above results (when We can obtain:

[0112] (11)

[0113] Equation (11) shows that after consistency mapping, the mean scale of the same feature dimension in the two operating modes reaches a statistical consistency, thus reflecting the result of the "scale correction / alignment". It should be noted that Equation (11) is a statistical description of the alignment effect, and its conclusion comes from the scaling mapping, not from the application of new constraints; the consistency mapping only adjusts the amplitude scale of each feature dimension and does not change the temporal evolution structure of the features. Figure 2 As shown, this is a schematic diagram illustrating the effect of data alignment through scaling of different feature dimensions.

[0114] Furthermore, let the total number of time windows used for analysis be... ,in, These are window numbers arranged in chronological order. The mapped feature vector for each window... Construct an insulation degradation feature matrix by splicing the data in chronological order. for:

[0115] (12)

[0116] in, The rows correspond to different feature dimensions, including ordinal component features and target harmonic amplitude features, while the columns correspond to sliding time window samples arranged in chronological order. By constructing the insulation degradation feature matrix, consistent feature inputs across operating modes can be provided to subsequent insulation health index construction, anomaly detection, trend prediction, or data-driven model training processes while maintaining the feature-time two-dimensional structure, thereby improving the stability and consistency of insulation degradation feature expression under cross-operating mode conditions.

[0117] In the constructed cross-operation mode consistent feature matrix Based on this, in order to highlight the key sequence component characteristics and target harmonic characteristics that are highly correlated with the insulation degradation mechanism, and to suppress the weakly correlated characteristics that are more sensitive to operating point fluctuations, this invention introduces a mechanism weighting coefficient vector. By performing weighted reconstruction on each row of the feature matrix, a structured insulation degradation feature matrix constrained by mechanism is obtained. .

[0118] Let the feature dimension be... The mechanism weight vector is then defined as: (13)

[0119] in, Indicates the first The sensitivity weight of a feature to the insulation degradation mechanism is determined by the weight of the feature. The larger the weight, the stronger the correlation between the feature and the insulation degradation mechanism.

[0120] Preferably, the mechanism weighting coefficient in this embodiment The weights are determined using a mechanistic prior method. Based on the mechanism by which insulation degradation leads to asymmetric and nonlinear enhancement of the three phases, higher priority is given to the asymmetric component characteristics and target harmonic characteristics. The weights are then suppressed by considering the stability of these characteristics in the operating data, thus obtaining weights that satisfy both the mechanism and engineering stability requirements.

[0121] Specifically, a mechanism priority coefficient is set for each feature dimension. Among these, the negative-sequence component characteristics and zero-sequence component characteristics have higher priority than the target harmonic amplitude characteristics. The priority coefficient of this mechanism is pre-set based on the strength of the correlation between different characteristics and the insulation degradation mechanism, and can be determined by combining prior mechanism analysis with historical sample statistical results. A stability suppression coefficient is also set. This reduces the weight of features that are more sensitive to fluctuations in the operating point. The stability suppression coefficient is preferably determined by the degree of dispersion of the feature, for example: (14)

[0122] in, For the first The coefficient of variation of the dimensional feature over the entire sliding time window. and These are the mean and standard deviation of this feature dimension, respectively. It is a small positive number.

[0123] Furthermore, construct the unnormalized quantity of the weights. for: (15)

[0124] This embodiment adopts and Multiplying is because the two represent the first and second parts respectively. n Two different dimensions of a feature: Indicates the relevance or priority of mechanisms. This indicates stability or disturbance resistance. The product of the two is... This can be understood as the comprehensive effective weight of features; that is, only features that are "strongly correlated with the insulation degradation mechanism" and "relatively stable under different operating conditions" will receive a higher weight. Because and All of them are dimensionless coefficients and have been normalized or constrained by intervals, so direct multiplication is reasonable and will not introduce dimensional issues.

[0125] And normalize it to obtain the mechanism weight coefficient: (16)

[0126] By using the above methods, features that are more strongly associated with the insulation degradation mechanism and more stable with changes in the operating point can be given greater weight, thereby enhancing the expressive power of mechanism-related features and suppressing the interference of unstable features on the results.

[0127] Based on the weight vector, the final weighted feature matrix is ​​constructed as follows: (17)

[0128] in, Indicates A diagonal matrix constructed for its diagonal elements; The obtained cross-operation mode consistent feature matrix; The structured insulation degradation characteristic matrix is ​​constrained by the mechanism, i.e., the final characteristic matrix.

[0129] It should be noted that, It is not an arbitrary intermediate process quantity, but a final feature expression form oriented towards engineering applications: while maintaining the "feature-time" two-dimensional structure, it simultaneously achieves cross-operation mode scale consistency and mechanism-sensitive feature enhancement. It can be directly used as a unified input for subsequent insulation status assessment and trend analysis, such as for constructing insulation health index, anomaly detection and early warning, or trend prediction model input, thereby meeting the engineering application requirements for stable and interpretable feature expression under complex operating conditions of pumped storage units.

[0130] Based on the obtained mechanism-constrained structured insulation degradation feature matrix Insulation status assessment and early warning output are performed. Let the total number of sliding time windows within the analysis period be... For any window ,Pick The The columns serve as the final feature representation of this window, and an insulation health index is constructed based on them. Preferably, the health index can be obtained by aggregating the features after normalization of each dimension, for example: (18)

[0131] in, for No. Line in window The value of ; For the normalization function, it is preferable to use minimum-maximum normalization or baseline normalization based on historical health data, so that... It falls within a preset range for threshold determination.

[0132] Furthermore, set early warning thresholds. ,when Output early warning information, when Output alarm information; and can be combined with The rate of change is used to determine the trend; for example, when multiple consecutive windows meet the following conditions... At that time, it was determined to be an accelerated deterioration and the warning level was upgraded, among which, This is the rate of change threshold. The output includes an insulation health index sequence. And the corresponding warning level information.

[0133] Another aspect of this embodiment also provides a system for constructing insulation degradation characteristics of generator motors, such as... Figure 3 As shown, the system includes:

[0134] The data acquisition module is used to simultaneously acquire the three-phase voltage signal and three-phase stator current signal of the pumped storage unit during stable operation, and to simultaneously acquire the operating status quantity of the pumped storage unit as an operating mode identification signal. The operating mode identification signal is used to indicate whether the pumped storage unit is in power generation mode or electric mode.

[0135] An insulation degradation feature vector construction module is used to preprocess the acquired three-phase stator current signal to obtain a preprocessed windowed three-phase stator current sequence, extract the zero-sequence component, positive-sequence component and negative-sequence component of the preprocessed windowed three-phase stator current sequence in each time window, and extract the harmonic features of the preprocessed windowed three-phase stator current sequence in each time window, thereby constructing an insulation degradation feature vector.

[0136] An insulation degradation feature sequence construction module is used to determine a mode sensitivity factor based on the insulation degradation feature vector, and use the mode sensitivity factor in combination with a consistency mapping function to perform scale unification on the insulation degradation feature vector to obtain a cross-mode consistent insulation degradation feature sequence. The mode sensitivity factor is the scale ratio of the electric mode to the power generation mode corresponding to each dimension feature in the insulation degradation feature vector.

[0137] The feature matrix determination module is used to construct a structured insulation degradation feature matrix based on the cross-mode consistent feature sequence, and to obtain the final feature matrix using a mechanism constraint method, thereby performing insulation status assessment and early warning output.

[0138] Other technical features of the generator motor insulation degradation feature construction system described in this embodiment are similar to the corresponding generator motor insulation degradation feature construction method, and will not be repeated here.

[0139] The steps of the various methods described above are only for clarity. In practice, they can be combined into one step or some steps can be split into multiple steps. As long as they include the same logical relationship, they are all within the scope of protection of this application. Adding insignificant modifications or introducing insignificant designs to the algorithm or process, but without changing the core design of the algorithm and process, are also within the scope of protection of this application.

[0140] Furthermore, some embodiments of this application also provide an electronic device. The electronic device can be various forms of digital computer, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, etc. The electronic device can also be various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices.

[0141] The electronic device includes: one or more processors; and a memory storing computer program instructions, which, when executed, cause the processor to perform a method for constructing insulation degradation characteristics of a generator motor as provided in any one or more of the above embodiments. The electronic device includes: one or more central processing units (CPUs), and interfaces for connecting various components, such as displays, infrared sensors, and cameras. That is, the various components are interconnected using different buses and can be mounted on a common motherboard or otherwise installed as needed. The processor can process instructions executed within the electronic device, including instructions stored in or on memory to display graphical information of a GUI on an external input / output device (such as a display device coupled to the interface). In some other embodiments, multiple processors and / or multiple buses can be used with multiple memories and multiple memory sets, if desired. Similarly, multiple electronic devices can be connected, each providing some of the necessary operations. The components, their connections and relationships, and their functions shown in this embodiment are merely examples and are not intended to limit the implementation of the present application described and / or claimed herein.

[0142] In a preferred embodiment of this invention, the electronic device may further include an input device and an output device. The processing unit, memory, input device, and output device may be connected via a bus or other means.

[0143] The input device can receive input numerical or character information, and generate key signal inputs related to user settings and function control of the electronic device, such as a touchscreen, keypad, mouse, trackpad, touchpad, joystick, one or more mouse buttons, trackball, joystick, etc. The output device may include a display device, auxiliary lighting devices (e.g., LEDs), and haptic feedback devices (e.g., vibration motors). The display device may include, but is not limited to, liquid crystal displays, light-emitting diode displays, and plasma displays. In some embodiments, the display device may be a touchscreen.

[0144] To provide interaction with the user, the electronic device can be a computer. The computer has: a display device (e.g., a cathode ray tube or LCD monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback); and input from the user can be received in any form (e.g., voice input or tactile input).

[0145] In this embodiment, a computer-readable medium stores a computer program / instructions, which, when executed by a processor, implement a method for constructing insulation degradation characteristics of a generator motor provided in any one or more of the above embodiments. This computer-readable medium may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into that device. The computer-readable medium carries one or more computer-readable instructions.

[0146] Memory can serve as a non-transitory computer-readable storage medium, used to store non-transitory software programs, non-transitory computer-executable programs, and modules. The central processing unit executes various server functions and data processing by running the non-transitory software programs, instructions, and modules stored in the memory, thereby implementing the program instructions / modules corresponding to the methods provided in any one or more of the embodiments described above in this application.

[0147] The memory may include a program storage area and a data storage area, wherein the program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the electronic device, etc. Furthermore, the memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory may optionally include memory remotely located relative to the processor, and these remote memories may be connected to the electronic device via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0148] It should be noted that the computer-readable medium described in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. Computer-readable media can be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to, electrical connections having one or more wires, portable computer disks, hard disks, random access memory, read-only memory, erasable programmable read-only memory, optical fibers, portable compact disk read-only memory, optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, a computer-readable medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0149] Computer-readable media include permanent and non-permanent, removable and non-removable media, which can store information by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory, static random access memory, dynamic random access memory, other types of random access memory, read-only memory, electrically erasable programmable read-only memory, flash memory or other memory technologies, read-only optical discs, digital versatile optical discs or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0150] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including local area networks (LANs) or wide area networks (WANs), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0151] In the above embodiments, all or part of the implementation can be achieved through software, hardware, firmware, or any combination thereof. For example, it can be implemented using an application-specific integrated circuit (ASIC), a general-purpose computer, or any other similar hardware device. In some embodiments, the software program of this application can be executed by a processor to implement the above steps or functions. Similarly, the software program of this application (including related data structures) can be stored in a computer-readable recording medium, such as RAM memory, magnetic or optical drives, floppy disks, and similar devices. In addition, some steps or functions of this application can be implemented in hardware, for example, as circuitry that cooperates with a processor to perform the various steps or functions.

[0152] The computer program product provided in this application includes one or more computer programs / instructions. When executed by a processor, these computer programs / instructions generate, in whole or in part, the processes or functions described in this application. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may 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, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium may 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.

[0153] The flowcharts or block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of devices, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-specific system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0154] The scope of this application is defined by the appended claims rather than the foregoing description, and is therefore intended to encompass all variations falling within the meaning and scope of equivalents of the claims. No reference numerals in the claims should be construed as limiting the scope of the claims. Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices recited in a device claim may also be implemented by a single unit or device in software or hardware. Terms such as "first," "second," etc., are used only for distinguishing descriptions and do not indicate any particular order, nor should they be construed as indicating or implying relative importance.

[0155] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily made by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims, and the above embodiments should be regarded as exemplary and non-limiting.

Claims

1. A method for constructing insulation degradation characteristics of a generator motor, characterized in that, The method includes: During the stable operation of the pumped storage unit, the three-phase voltage signal and three-phase stator current signal of the pumped storage unit are collected synchronously, and the operating status of the pumped storage unit is collected synchronously as an operating mode identification signal. The operating mode identification signal is used to indicate whether the pumped storage unit is in power generation mode or electric mode. After preprocessing the acquired three-phase stator current signal, a preprocessed windowed three-phase stator current sequence is obtained. The zero-sequence component, positive-sequence component, and negative-sequence component of the preprocessed windowed three-phase stator current sequence in each time window are extracted, and the harmonic features of several orders of the preprocessed windowed three-phase stator current sequence in each time window are extracted to construct an insulation degradation feature vector. Based on the insulation degradation feature vector, a mode sensitivity factor is determined. The mode sensitivity factor is then combined with a consistency mapping function to perform scale unification on the insulation degradation feature vector to obtain a cross-mode consistent insulation degradation feature sequence. The mode sensitivity factor is used to characterize the scale ratio of the electric mode to the power generation mode corresponding to each dimension feature in the insulation degradation feature vector. Based on the cross-mode consistent insulation degradation feature sequence, a structured insulation degradation feature matrix is ​​constructed, and the final feature matrix is ​​obtained by using a mechanism constraint method, thereby performing insulation status assessment and early warning output.

2. The method for constructing insulation degradation characteristics of a generator motor according to claim 1, characterized in that, The step of extracting harmonic features of several orders of the preprocessed windowed three-phase stator current sequence in each time window and then constructing an insulation degradation feature vector includes: using Fast Fourier Transform to extract the amplitude of several orders of harmonic features of the preprocessed windowed three-phase stator current sequence in each time window, and determining the target harmonic order set based on the set mechanism prior constraint method; and concatenating the amplitudes of the negative sequence component, the zero sequence component, and the harmonic features of the determined order to construct the insulation degradation feature vector.

3. The method for constructing insulation degradation characteristics of a generator motor according to claim 2, characterized in that, The method for determining the target harmonic order set based on the set mechanistic prior constraints includes: A candidate order set is constructed based on the symmetry destruction and dielectric nonlinear enhancement mechanism caused by insulation degradation. The symmetry destruction corresponds to the enhancement of negative or zero-sequence components and is accompanied by changes in even-order harmonic components. The dielectric nonlinear enhancement corresponds to changes in several odd-order harmonic components. Based on the statistical screening of each candidate order in the candidate order set using sample windows, the harmonic amplitude sequence corresponding to the sample window is divided into a healthy segment and a suspected degradation segment. The discrimination index and trend correlation index are used to determine the discrimination and degradation trend correlation of any candidate order in the healthy segment and the suspected degradation segment, respectively. The healthy segment is selected from the window set where the insulation status is confirmed to be normal after maintenance or in the early stage of operation, or the window set where the negative sequence component, zero sequence component and harmonic level are within the baseline threshold range and there are no abnormal alarms. The suspected degradation segment is selected from the window set where there are offline detection or maintenance records indicating degradation, or the window set that meets the degradation criteria. The current comprehensive score is determined based on the correlation between discrimination and degradation trend, and candidate orders with scores below the first threshold are eliminated. Calculate the correlation between the amplitude of the harmonic characteristics of the current order and the load characterization quantity, and denot it as characterization correlation. If the characterization correlation is higher than the second threshold and the comprehensive score of the corresponding order is lower than the first threshold, then remove the order from the candidate order set. To ensure the stability of the selected harmonic order under different operating conditions, the comprehensive score and the characterization correlation are repeatedly calculated in power generation mode, electric mode, and different load ranges to obtain the final target harmonic order set.

4. The method for constructing insulation degradation characteristics of a generator motor according to claim 2, characterized in that, The determination of the mode sensitivity factor based on the insulation degradation feature vector includes: Determine the dimension of the insulation degradation feature vector This dimension is obtained by adding the zero-sequence component, the negative-sequence component, and the order of the target harmonic; Based on the operating mode identification signal, all sliding time windows are divided into a power generation mode window set and a motoring mode window set. The statistical scaling parameter for each feature dimension is calculated for both the power generation mode window set and the motoring mode window set, and is expressed as follows: ;in, Indicates the first Each dimension corresponds to the mean scale of the feature under the power generation mode. Indicates the first The mean scale of the dimensional feature in the electric mode. These represent the number of windows within the power generation mode window set and the electric mode window set, respectively. Indicates the sliding time window number. For the first Feature vectors corresponding to each dimension; Definition of the first The pattern sensitivity factor corresponding to each dimension of the feature is represented as: ; in, Indicates the first The scale ratio of the features corresponding to each dimension in electric mode relative to power generation mode; It is a tiny positive number that approaches zero, used to prevent the denominator from being zero.

5. The method for constructing insulation degradation characteristics of a generator motor according to claim 4, characterized in that, The step of using the mode sensitivity factor combined with the consistency mapping function to scale the insulation degradation feature vector to obtain a cross-mode consistent insulation degradation feature sequence includes: Using a consistent mapping function to represent the first The feature corresponding to the dimension in the th dimension is at the th . The values ​​within each sliding time window are: Thus, a cross-mode consistent insulation degradation characteristic sequence is obtained. The consistency mapping function is used to mitigate the systematic scale shift caused by the difference between the power generation mode and the electric mode. It only adjusts the amplitude scale of each feature dimension to achieve scale alignment.

6. The method for constructing insulation degradation characteristics of a generator motor according to claim 5, characterized in that, Based on the cross-mode consistent insulation degradation feature sequence, a structured insulation degradation feature matrix is ​​constructed, including: Verify the alignment effect of the consistency mapping on the scale shift caused by the difference between power generation and electric modes, i.e., for any feature dimension. ,Depend on And when hour, get: ; Therefore, after the consistency mapping, the mean scale of the same feature dimension in the two operating modes becomes consistent in a statistical sense; Let the total number of time windows used for analysis be . ,in, The window numbers are arranged in chronological order, and the mapped feature vectors for each window are... Construct an insulation degradation feature matrix by splicing the data in chronological order. ,for: ; in, The rows correspond to different feature dimensions, and the columns correspond to sliding time window samples arranged in chronological order.

7. The method for constructing insulation degradation characteristics of a generator motor according to claim 6, characterized in that, The mechanism constraint method includes: Set a mechanism priority coefficient for each dimension's corresponding feature. Among them, the priority coefficients corresponding to the negative sequence component characteristics and the zero sequence component characteristics are higher than the priority coefficients corresponding to the amplitude characteristics of the target harmonic. Set stability suppression coefficient The stability suppression coefficient is determined by the degree of dispersion of the feature; The sensitivity weights of each dimension are obtained by multiplying the priority coefficient and the stability inhibition coefficient by the computer and normalizing the product. Dimensions based on insulation degradation feature vectors Define the mechanism weight vector as follows: ,in, Indicates the first Each dimension corresponds to a feature's sensitivity weight to the insulation degradation mechanism; the larger the weight, the stronger the correlation between the feature and the insulation degradation mechanism. Based on the aforementioned mechanism weight vector, the weighted final feature matrix is ​​constructed as follows: ;in, Represents the mechanism weight vector A diagonal matrix constructed for the diagonal elements.

8. The method for constructing insulation degradation characteristics of a generator motor according to claim 7, characterized in that, The stability suppression coefficient is expressed as: ;in, For the first Each dimension corresponds to the coefficient of variation of the feature over the entire sliding time window. These are the mean and standard deviation of this feature dimension, respectively. It is a tiny positive number that approaches zero.

9. A system for constructing insulation degradation characteristics of a generator motor, characterized in that, The system includes: The data acquisition module is used to simultaneously acquire the three-phase voltage signal and three-phase stator current signal of the pumped storage unit during stable operation, and to simultaneously acquire the operating status quantity of the pumped storage unit as an operating mode identification signal. The operating mode identification signal is used to indicate whether the pumped storage unit is in power generation mode or electric mode. An insulation degradation feature vector construction module is used to preprocess the acquired three-phase stator current signal to obtain a preprocessed windowed three-phase stator current sequence, extract the zero-sequence component, positive-sequence component and negative-sequence component of the preprocessed windowed three-phase stator current sequence in each time window, and extract the harmonic features of the preprocessed windowed three-phase stator current sequence in each time window to construct an insulation degradation feature vector. An insulation degradation feature sequence construction module is used to determine a mode sensitivity factor based on the insulation degradation feature vector, and use the mode sensitivity factor in combination with a consistency mapping function to perform scale unification on the insulation degradation feature vector to obtain a cross-mode consistent insulation degradation feature sequence. The mode sensitivity factor is used to characterize the scale ratio of the electric mode to the power generation mode corresponding to each dimension feature in the insulation degradation feature vector. The feature matrix determination module is used to construct a structured insulation degradation feature matrix based on the cross-mode consistent insulation degradation feature sequence, and to obtain the final feature matrix using a mechanism constraint method, thereby performing insulation status assessment and early warning output.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1 to 8.

Images