Hydro-generator ultrahigh frequency partial discharge identification method and device based on data driving

Abstract

The invention provides a hydro-generator ultrahigh frequency partial discharge identification method and device based on data driving, and belongs to the technical field of partial discharge monitoring. The method comprises the following steps: acquiring a monitoring signal acquired by a coupling device; discretizing the monitoring signal to obtain an amplitude sequence; inputting the amplitude sequence into a pre-constructed signal diagnosis model, and generating a diagnosis result indicating whether the hydraulic generator works normally; if the diagnosis result shows that the working state of the hydro-generator is abnormal, extracting a first feature of the monitoring signal; if the interference signal corresponding to the first feature is not found in the constructed background noise library, extracting a second feature of the monitoring signal; and matching a partial discharge event corresponding to the second feature from a constructed partial discharge case library by using the second feature, the first feature including an energy feature, and the second feature including a time domain and / or frequency domain feature. According to the method, data driving, a background noise library and a partial discharge case library are combined, and rapid and accurate identification of partial discharge signals is realized.

Description

Technical Field

[0001] The present application belongs to the technical field of partial discharge monitoring, and specifically relates to a data-driven ultra-high frequency partial discharge identification method for a hydro-turbine generator, a data-driven ultra-high frequency partial discharge identification device for a hydro-turbine generator, a computer device, and a machine-readable storage medium. Background Art

[0002] As the core equipment of a hydropower station, the health of the insulation system of a hydro-turbine generator is directly related to the safety and stability of the entire power system. In actual production, the detection of partial discharge (PD) signals is an important means to reflect the insulation status of large-scale power equipment. With the continuous expansion of hydropower installed capacity, higher requirements are placed on the monitoring and evaluation of the insulation status of hydro-turbine generators. In recent years, compared with low-frequency PD monitoring technology, ultra-high frequency PD monitoring technology has been gradually adopted due to its higher sensitivity and better anti-interference performance. However, due to the complex operating conditions of hydro-turbine generators, the interference of partial discharge signals monitored by ultra-high frequency PD is still very serious. How to achieve accurate PD signal recognition is a technical difficulty in the field of ultra-high frequency PD monitoring of hydro-turbine generators. Summary of the invention

[0003] In view of this, the purpose of the embodiments of the present application is to provide a data-driven ultra-high frequency partial discharge identification method for a hydro-turbine generator, a data-driven ultra-high frequency partial discharge identification device for a hydro-turbine generator, a computer device and a machine-readable storage medium, so as to overcome the technical problem in the prior art that it is difficult to accurately identify partial discharge signals due to the complex operating conditions of the hydro-turbine generator.

[0004] To achieve the above-mentioned purpose, the first aspect of the present application provides a data-driven method for identifying ultra-high frequency partial discharge of a hydro-turbine generator, the method comprising: Obtaining the monitoring signal collected by the ultra-high frequency partial discharge coupling device in the current monitoring period; Discretizing the monitoring signal to obtain an amplitude sequence; Inputting the amplitude sequence into a pre-built signal diagnosis model to generate a diagnosis result of whether the hydro-generator is working normally; If the diagnosis result indicates that the working state of the hydro-generator is abnormal, extracting the first feature of the monitoring signal; If no interference signal corresponding to the first feature is found in the constructed background noise library, extracting the second feature of the monitoring signal; Using the second feature, matching a partial discharge event corresponding to the second feature from a constructed partial discharge case library; Among them, each interference signal has its own corresponding first feature, and each partial discharge event has its own corresponding second feature. The first feature includes energy features, and the second feature includes time domain and / or frequency domain features. The signal diagnosis model is obtained by training the classification network and optimizing parameters.

[0005] In a specific embodiment of the present application, the signal diagnosis model is obtained after training and optimizing the parameters of the classification network, including: Performing a first-stage training on the classification network using a first training set to obtain a first diagnostic model, wherein the first training set is composed of an amplitude sequence obtained by discretizing historical normal working state signals of a hydro-generator; Performing a second stage of training on the first diagnostic model using a second training set to obtain a second diagnostic model, wherein the second training set is composed of a specific amplitude sequence, and the historical monitoring signal corresponding to the specific amplitude sequence is an abnormal working state signal of a hydro-generator; Performing a third stage of training on the second diagnostic model using a third training set, wherein the third training set consists of a target amplitude sequence, and the target amplitude sequence is a historical amplitude sequence that meets a preset condition; Among them, the preset condition is: the diagnosis result obtained by inputting the second diagnostic model is that the working state of the hydro-generator is abnormal, and then it is manually determined that the monitoring signal corresponding to the input amplitude sequence is an abnormal working state signal of the hydro-generator.

[0006] In a specific embodiment of the present application, the types of classification networks include XGBoost, random forest, and extreme learning machine.

[0007] In a specific embodiment of the present application, the energy characteristic is power spectral density.

[0008] In a specific embodiment of the present application, the time domain feature includes at least one of a phase, an amplitude, a waveform feature, and an amplitude statistical feature.

[0009] In a specific embodiment of the present application, the waveform feature includes one of a rise time, a fall time, a duration and a pulse width.

[0010] In a specific embodiment of the present application, the types of interference signals in the background noise library include corona discharge generated from the outside of the hydro-generator, electromagnetic interference, pulse interference, interference signals caused by poor grounding during monitoring signal collection, electronic noise of the collection equipment itself during monitoring signal collection, environmental noise, erroneous signals generated by the collection equipment caused by temperature changes during monitoring signal collection, false discharge signals caused by humidity changes, signal transmission interference and human interference.

[0011] The second aspect of the present application provides a data-driven ultra-high frequency partial discharge identification device for a hydro-turbine generator, the device comprising: An acquisition module, used to obtain the monitoring signal collected by the ultra-high frequency partial discharge coupling device in the current monitoring period; A signal processing module, used for discretizing the monitoring signal to obtain an amplitude sequence; A diagnosis module, used for inputting the amplitude sequence into a pre-built signal diagnosis model to generate a diagnosis result of whether the hydro-generator is working normally; A first extraction module, used for extracting a first feature of the monitoring signal when the diagnosis result shows that the working state of the hydro-generator is abnormal; A second extraction module, configured to search for an interference signal corresponding to the first feature in the constructed background noise library, and if no interference signal corresponding to the first feature is found, extract a second feature of the monitoring signal; a signal recognition module, configured to use the second feature to match a partial discharge event corresponding to the second feature from a constructed partial discharge case library; Among them, each interference signal has its own corresponding first feature, and each partial discharge event has its own corresponding second feature. The first feature includes energy features, and the second feature includes time domain and / or frequency domain features. The signal diagnosis model is obtained by training the classification network and optimizing parameters.

[0012] A third aspect of the present application provides a computer device, comprising: a memory configured to store instructions; and The processor is configured to call the instruction from the memory and implement the data-driven ultra-high frequency partial discharge identification method of a hydro-turbine generator according to the first aspect of the present application when executing the instruction.

[0013] The fourth aspect of the present application provides a machine-readable storage medium having instructions stored thereon, the instructions being used to enable a machine to execute the data-driven method for identifying ultra-high frequency partial discharge of a hydro-turbine generator according to the first aspect of the present application.

[0014] In the above technical solution, a data-driven signal diagnosis model is used to diagnose the working state of the hydro-generator, and then the background noise library and the partial discharge case library are combined to distinguish interference signals from partial discharge events, thereby realizing rapid and accurate identification of ultra-high frequency partial discharge signals.

[0015] Other features and advantages of the embodiments of the present application will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present application, but do not constitute a limitation on the embodiments of the present application. In the accompanying drawings: Figure 1 A flow chart of a method for identifying ultra-high frequency partial discharge of a hydro-turbine generator based on data drive according to an embodiment of the present application is schematically shown; Figure 2 A block diagram schematically shows a composition of a data-driven hydro-generator ultra-high frequency partial discharge identification device according to an embodiment of the present application; Figure 3 The structural block diagram of a computer device according to an embodiment of the present application is schematically shown. DETAILED DESCRIPTION

[0017] The specific implementation of the embodiment of the present application is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present application, and is not used to limit the embodiment of the present application.

[0018] If there are descriptions involving "first", "second", etc. in the embodiments of the present application, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0019] In the current power system, due to the attenuation of ultra-high frequency partial discharge signals, it is difficult to obtain the true value of partial discharge signals through quantitative analysis. In current engineering applications, ultra-high frequency partial discharge monitoring usually uses a method that combines single-machine trend comparison and partial discharge case library for judgment. Its essence is a qualitative partial discharge value determination method. This qualitative analysis method has high requirements on the ability of on-site technicians. Therefore, the existing online partial discharge monitoring methods for hydropower generators have certain limitations.

[0020] In order to overcome the above defects, the embodiment of the present application proposes a data-driven method for identifying ultra-high frequency partial discharge of a hydro-turbine generator, the method comprising: Obtaining the monitoring signal collected by the ultra-high frequency partial discharge coupling device in the current monitoring period; Preprocess and discretize the collected monitoring signals to obtain an amplitude sequence; The amplitude sequence is input into a pre-built signal diagnosis model to generate a diagnosis result of whether the hydro-generator is working properly, wherein the signal diagnosis model is obtained by training the classification network and optimizing the parameters; If the diagnosis result indicates that the working state of the hydro-generator is abnormal, extracting a first feature of the monitoring signal, the first feature at least including an energy feature; If no interference signal corresponding to the first feature is found in the constructed background noise library, a second feature of the monitoring signal is extracted, wherein each interference signal has its own corresponding first feature, and the second feature includes at least a time domain feature and / or a frequency domain feature; The second feature of the monitoring signal is used to match a partial discharge event corresponding to the second feature from a constructed partial discharge case library, wherein each partial discharge event has its own corresponding second feature.

[0021] It should be understood that in the present application, the collected monitoring signals are discretized to obtain a pulse sequence, and the amplitude sequence is a sequence composed of the amplitudes of each pulse in the pulse sequence. In addition, the partial discharge signals corresponding to different types of partial discharge events are differentiated in the time domain and / or frequency domain, specifically referring to: the amplitude, frequency and waveform of the partial discharge signal change with the change of the partial discharge position and the partial discharge intensity. Therefore, as mentioned above, the second feature refers to these features that can reflect the differences in the partial discharge signal in the time domain or frequency domain. As mentioned above, the specific time domain features and / or frequency domain features used in the embodiment are not limited, and the feasible time domain features and / or frequency domain features need to be determined in combination with the specific type of turbine generator unit in the hydropower station.

[0022] The above embodiment of the present application combines a data-driven signal diagnosis model, a background noise library, and a partial discharge case library. First, the data-driven signal diagnosis model is used to identify whether the monitoring signal collected in the current monitoring period is a normal working state signal of the hydro-turbine generator or an abnormal working state signal of the hydro-turbine generator. The signal diagnosis model itself has the high-precision characteristics of a machine learning model. When the abnormal working state of the hydro-turbine generator is identified, the background noise library and the partial discharge case library are combined to determine whether the current monitoring signal is an interference signal or a partial discharge signal. If the diagnosis result of the signal diagnosis model is that the working state of the hydro-turbine generator is normal, then the monitoring signal corresponding to the amplitude sequence of the input signal diagnosis model is the normal working state signal of the hydro-turbine generator, because the monitoring signal The signal is analyzed only after being pre-processed by filtering, denoising, cleaning, etc., but there are still interferences that cannot be filtered out, such as periodic pulse interference. If the diagnosis result of the signal diagnosis model is that the working state of the hydro-turbine generator is abnormal, then the monitoring signal corresponding to the amplitude sequence of the input signal diagnosis model is the abnormal working state signal of the hydro-turbine generator. The abnormal working state signal of the hydro-turbine generator can be an industrial frequency cycle carrying interference signals such as periodic pulse interference, or an industrial frequency cycle carrying partial discharge pulses. There are significant differences in energy characteristics between interference signals and partial discharge pulses. Therefore, the interference signal and the partial discharge signal are distinguished by a first feature including at least an energy feature. After the partial discharge signal is identified, the partial discharge event type corresponding to the partial discharge signal is identified by a second feature.

[0023] In a comparative example, an unsupervised clustering method is used to implement an adaptive threshold, an adaptive threshold is used to distinguish interference and partial discharge signals, and an unsupervised clustering method is used to classify partial discharge signals. Because the types of interference signals and partial discharge signals are relatively large, there are more cluster centers when using the unsupervised clustering method, which makes the calculation amount relatively large, and the recognition efficiency is lower than the recognition method in this application that combines a data-driven signal diagnosis model, a background noise library, and a partial discharge case library.

[0024] Based on this, the embodiments of the present application as described above achieve rapid and accurate identification of ultra-high frequency partial discharge signals.

[0025] Figure 1 The data-driven method for identifying ultra-high frequency partial discharge of a hydro-turbine generator according to an embodiment of the present application is schematically shown. Figure 1 As shown, the data-driven method for identifying ultra-high frequency partial discharge of a hydro-turbine generator may include the following steps 102 to 112. It can be understood that the data-driven method for identifying ultra-high frequency partial discharge of a hydro-turbine generator may include all steps from step 102 to step 112, or may only include some of the steps.

[0026] Step 102, obtaining the monitoring signal collected by the ultra-high frequency partial discharge coupling device in the current monitoring period.

[0027] In the present application, the UHF partial discharge coupling device may be an UHF partial discharge sensor, an UHF partial discharge antenna, or a coupler composed of a transformer and a coupling capacitor.

[0028] Step 104: preprocess and discretize the collected monitoring signal to obtain an amplitude sequence. It is known that the preprocessing may include bandpass filtering, normalization processing, etc.

[0029] For example, the normalization process is performed as follows: ; in, x(t) represents the monitoring signal after bandpass filtering, max(|x(t)|)is the signal x(t) The maximum absolute amplitude of the normalized signal x b (t) The range is [-1,1].

[0030] Step 106, inputting the amplitude sequence obtained in step 104 into a pre-constructed signal diagnosis model to generate a diagnosis result of whether the hydro-generator is working normally.

[0031] Specifically, in the present application, the signal diagnosis model is obtained by training the classification network and optimizing the parameters.

[0032] As an example, the type of classification network may be XGBoost, random forest, extreme learning machine, RBF network, etc.

[0033] As an example, the signal diagnosis model is trained using the cross entropy loss function and regularization is used to prevent the model from overfitting.

[0034] It is known that in actual production, the amount of signal data collected by the ultra-high frequency partial discharge online monitoring method is huge, and most of them are signals of the normal working state of the hydro-generator. Therefore, there is a lack of abundant partial discharge signal samples that can be used to train the classification network. Therefore, based on the training method in the ordinary embodiment, the model is prone to problems such as overfitting and poor generalization ability. Therefore, as an improvement of the above-mentioned embodiment of the present application, the present application proposes a signal diagnosis model construction method based on multi-stage training. Accordingly, the signal diagnosis model is obtained after the classification network is trained and the parameters are optimized, which specifically includes the following steps: Step A1, using the first training set to perform the first stage training on the classification network to obtain the first diagnostic model, wherein the first training set is composed of an amplitude sequence obtained by discretizing the normal working state signal of the hydro-generator.

[0035] In this application, the samples used in the first stage of training can be called positive samples.

[0036] Step A2, using the second training set to perform the second stage training on the first diagnostic model to obtain the second diagnostic model, the second training set is composed of a specific amplitude sequence, and the historical monitoring signal corresponding to the specific amplitude sequence is an abnormal working state signal of the hydro-generator.

[0037] In this application, the samples used in the second stage of training can be called negative samples.

[0038] Step A3, performing a third stage of training on the second diagnostic model using a third training set, wherein the third training set consists of a target amplitude sequence, and the target amplitude sequence is a historical amplitude sequence that meets a preset condition.

[0039] Specifically, the preset condition described in the present application means that the diagnostic result obtained by inputting the second diagnostic model is that the working state of the hydro-generator is abnormal, and then the monitoring signal corresponding to the input amplitude sequence is manually determined to be the abnormal working state signal of the hydro-generator. The samples used in the third stage of training are samples obtained after being identified and manually marked by the second diagnostic model, and are also negative samples.

[0040] In the above improved embodiment, through multi-stage training, full use is made of positive samples and negative samples, and the third training set is an expanded negative sample obtained by identification and manual labeling of the second diagnostic model. The efficiency of sample expansion is higher than that of manually labeling negative samples in a large number of historical monitoring signals directly in the ordinary embodiment. Therefore, the construction efficiency of the signal diagnosis model and the recognition accuracy of the constructed signal diagnosis model are improved through multi-stage training.

[0041] In summary, the above improved embodiments improve the efficiency and accuracy of partial discharge identification.

[0042] Step 108, if the diagnosis result shows that the working state of the hydro-generator is abnormal, extract the first feature of the monitoring signal; if the diagnosis result shows that the working state of the hydro-generator is normal, the partial discharge identification ends.

[0043] In the present application, the first feature at least includes an energy feature. For example, the energy feature is a power spectrum density.

[0044] Step 110: If the interference signal corresponding to the first feature of the extracted monitoring signal is not found in the constructed background noise library, the second feature of the monitoring signal is extracted, wherein each interference signal has its own corresponding first feature; if the interference signal corresponding to the first feature of the extracted monitoring signal is found in the constructed background noise library, the monitoring signal collected in the current monitoring period is determined to be the interference signal, and the partial discharge identification is completed.

[0045] In the present application, the second feature includes at least a time domain feature and / or a frequency domain feature.

[0046] In an optional embodiment of the present application, the time domain features include one or more of phase, amplitude, waveform features and amplitude statistical features. The waveform features include one or more of rise time, fall time, duration and pulse width. The amplitude statistical features include one or more of amplitude mean value and amplitude standard deviation.

[0047] Because different types of partial discharge signals correspond to different frequency spectrum distributions, in an optional embodiment of the present application, the frequency domain feature may be the intensity, frequency distribution, etc. of different frequency components in the spectrum.

[0048] As an example, the types of interference signals in the background noise library include corona discharge generated from the outside of the hydro-generator, electromagnetic interference, pulse interference, interference signals caused by poor grounding during monitoring signal acquisition, electronic noise of the acquisition equipment itself during monitoring signal acquisition, environmental noise, false signals generated by the acquisition equipment due to temperature changes during monitoring signal acquisition, false discharge signals caused by humidity changes, signal transmission interference, and human interference, etc. The sources of these interference signals are shown in Table 1 below.

[0049] Table 1

[0050] Step 112, using the second feature of the extracted monitoring signal to match the partial discharge event corresponding to the second feature from the constructed partial discharge case library, wherein each partial discharge event has its own corresponding second feature; if the partial discharge event corresponding to the second feature is not matched from the constructed partial discharge case library using the second feature of the extracted monitoring signal, the monitoring signal is manually marked, and if it is manually determined that the difference between the first feature of the monitoring signal and the first feature of the interference signal in the background noise library is within a first preset range, the type of interference signal to which the monitoring signal belongs is manually marked, otherwise, the type of partial discharge event to which the monitoring signal belongs is manually marked. When manually analyzing and determining a new partial discharge event type, the result can be conducted in an offline partial discharge test mode, for example, a partial discharge test is conducted at each ultra-high frequency partial discharge coupling device layout position of the stator winding of the hydro-turbine generator in an offline test mode, the partial discharge signals collected at each layout position are collected, and the characteristic parameters of these partial discharge signals are extracted, and the partial discharge event type of the monitoring signal to be currently classified is determined by matching and analyzing the extracted characteristic parameters with the characteristic parameters of the monitoring signal to be currently classified. Preferably, before extracting the characteristic parameters of these partial discharge signals, the partial discharge signals collected at each distribution point are corrected using the maximum attenuation coefficient, and the maximum attenuation coefficient can be determined through the process in the common embodiment.

[0051] For example, by injecting a test partial discharge signal at each preset injection point of the stator winding bar, and then collecting the partial discharge signal at the location of the ultra-high frequency partial discharge coupling device, and then determining the transfer function and attenuation coefficient of the stator winding bar through a network vector analyzer, etc., thereby determining the maximum attenuation coefficient, the partial discharge signal collected at each location is reconstructed through the transfer function, and the reconstructed partial discharge signal is multiplied by the maximum attenuation coefficient to complete the correction of the partial discharge signal.

[0052] It should be understood that the types of interference signals in the background noise library are gradually improved, and the types of partial discharge events in the partial discharge case library are also gradually improved. When the monitoring signals cannot be classified using the current background noise library and partial discharge case library, it is necessary to classify the monitoring signals through manual analysis, and add the new interference signal types obtained by classification to the background noise library or add the new partial discharge events obtained by classification to the partial discharge case library.

[0053] As an example, the partial discharge case library includes a variety of typical partial discharge events and a variety of common partial discharge events. Typical partial discharge events are partial discharge events when the turbine generator is under extreme conditions, and partial discharge events when the turbine generator is under normal conditions are common partial discharge events. A typical partial discharge event is a partial discharge case under extreme humidity conditions, specifically referring to: simulating partial discharge under high humidity environment, because moisture can significantly change the performance of insulating materials and increase the probability of partial discharge. Through partial discharge simulation under such conditions, partial discharge events under such conditions are obtained. Another typical partial discharge event is a partial discharge case under high temperature environment, specifically referring to: partial discharge simulation under conditions higher than normal operating temperature, because high temperature may cause accelerated aging of insulating materials, thereby causing partial discharge. Through partial discharge simulation under such conditions, partial discharge events under such conditions are obtained. Another typical partial discharge event is a partial discharge case under strong electromagnetic interference environment, specifically referring to: simulating partial discharge under an environment with strong electromagnetic interference. In specific applications, the complex electromagnetic interference environment in the actual power system is simulated to obtain partial discharge events under such conditions. Another typical partial discharge event is the partial discharge case after the aging of the insulating material, which specifically refers to: using the aged insulating material or accelerated aging experiment to simulate the long-term operation effect of the hydro-turbine generator, investigating the influence of insulation aging on partial discharge, and obtaining the partial discharge event under this condition. Another typical partial discharge event is the partial discharge case caused by the hydro-turbine generator under mechanical stress, which specifically refers to: during the partial discharge test of the hydro-turbine generator, mechanical vibration or impact is applied to the hydro-turbine generator, because mechanical stress or impact may cause damage to the insulation structure, thereby generating partial discharge. By investigating the partial discharge caused by such insulation structure damage, the partial discharge event under this condition is obtained. Another typical partial discharge event is the partial discharge case caused by different types of insulation defects, which specifically refers to: constructing insulation defect models with different types and sizes, such as bubbles, cracks, impurities, etc., investigating the influence of these different insulation defects on partial discharge, and obtaining the partial discharge event under this condition. Another typical partial discharge event is the partial discharge case of the turbine generator under extreme operating conditions, which specifically refers to: simulating the extreme operating conditions of the turbine generator, such as rapid voltage increase, rapid voltage reduction and overload, because these extreme operating conditions may lead to aggravation of partial discharge phenomenon, and through the partial discharge simulation under these extreme operating conditions, the partial discharge events under such conditions are obtained. Another typical partial discharge event is the partial discharge case under compound fault conditions, which specifically refers to: conducting partial discharge simulation tests on the turbine generator under multiple extreme conditions at the same time, such as the simultaneous existence of high temperature, high humidity and mechanical stress, and examining the comprehensive impact of compound fault conditions on partial discharge, and obtaining partial discharge events under such conditions.

[0054] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0055] Corresponding to the data-driven method for identifying ultra-high frequency partial discharge of a hydro-turbine generator in the above embodiment, Figure 2 The block diagram of the composition of the data-driven ultra-high frequency partial discharge identification device 400 for a hydro-turbine generator provided in an embodiment of the present application is schematically shown. For the sake of convenience of explanation, only the parts related to the embodiment of the present application are shown.

[0056] like Figure 2 As shown, the data-driven ultra-high frequency partial discharge identification device 400 for a hydro-turbine generator includes: The acquisition module 410 is used to obtain the monitoring signal collected by the ultra-high frequency partial discharge coupling device in the current monitoring period; The signal processing module 420 is used to discretize the monitoring signal to obtain an amplitude sequence; A diagnosis module 430, for inputting the amplitude sequence into a pre-built signal diagnosis model to generate a diagnosis result of whether the hydro-generator is working normally; A first extraction module 440 is used to extract the first feature of the monitoring signal when the diagnosis result shows that the working state of the hydro-generator is abnormal; A second extraction module 450 is used to search for an interference signal corresponding to the first feature in the constructed background noise library, and if no interference signal corresponding to the first feature is found, extract the second feature of the monitoring signal; A signal identification module 460, configured to use the second feature to match a partial discharge event corresponding to the second feature from a constructed partial discharge case library; Among them, each interference signal has its own corresponding first feature, and each partial discharge event has its own corresponding second feature. The first feature includes energy features, and the second feature includes time domain and / or frequency domain features. The signal diagnosis model is obtained by training the classification network and optimizing parameters.

[0057] The process of each module in the data-driven hydro-generator ultra-high frequency partial discharge identification device 400 provided in the embodiment of the present application realizing its respective function can be specifically referred to the aforementioned Figure 1 The description of the illustrated embodiment and other related method embodiments will not be repeated here.

[0058] It should be noted that the information interaction, execution process and other contents between the above modules are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0059] Figure 3 The structure block diagram of the computer device according to the embodiment of the present application is schematically shown. In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 3 As shown. The computer device includes a processor A01, a network interface A02, a display screen A04, an input device A05 and a memory (not shown in the figure) connected through a system bus. Among them, the processor A01 of the computer device is used to provide computing and control capabilities. The memory of the computer device includes an internal memory A03 and a non-volatile storage medium A06. The non-volatile storage medium A06 stores an operating system B01 and a computer program B02. The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 in the non-volatile storage medium A06. The network interface A02 of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor A01, a method for identifying ultra-high frequency partial discharge of a hydro-turbine generator based on data driving is implemented. The display screen A04 of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device A05 of the computer device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0060] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0061] In one embodiment, the data-driven hydro-turbine generator ultra-high frequency partial discharge identification device 400 provided in the present application can be implemented in the form of a computer program. The computer program can be used in Figure 3 The computer device shown in the figure is run. The memory of the computer device can store various program modules constituting the data-driven hydro-generator ultra-high frequency partial discharge identification device 400. The computer program composed of various program modules enables the processor to execute the steps of the data-driven hydro-generator ultra-high frequency partial discharge identification method in various embodiments of the present application described in this specification.

[0062] In one embodiment, the present application further provides a machine-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the data-driven ultra-high frequency partial discharge identification method for a hydro-turbine generator in the above embodiment is implemented.

[0063] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0064] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0065] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0066] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A data-driven method for identifying ultra-high frequency partial discharge of a hydro-turbine generator, characterized in that: include: Obtaining the monitoring signal collected by the ultra-high frequency partial discharge coupling device in the current monitoring period; Discretizing the monitoring signal to obtain an amplitude sequence; Inputting the amplitude sequence into a pre-built signal diagnosis model to generate a diagnosis result of whether the hydro-generator is working normally; If the diagnosis result indicates that the working state of the hydro-generator is abnormal, extracting the first feature of the monitoring signal; If no interference signal corresponding to the first feature is found in the constructed background noise library, extracting the second feature of the monitoring signal; Using the second feature, matching a partial discharge event corresponding to the second feature from a constructed partial discharge case library; Among them, each interference signal has its own corresponding first feature, and each partial discharge event has its own corresponding second feature. The first feature includes energy features, and the second feature includes time domain and / or frequency domain features. The signal diagnosis model is obtained by training the classification network and optimizing parameters.

2. The data-driven method for identifying ultra-high frequency partial discharge of a hydro-turbine generator according to claim 1 is characterized in that: After training and optimizing the parameters of the classification network, the signal diagnosis model is obtained, including: Performing a first-stage training on the classification network using a first training set to obtain a first diagnostic model, wherein the first training set is composed of an amplitude sequence obtained by discretizing historical normal working state signals of a hydro-generator; Performing a second stage of training on the first diagnostic model using a second training set to obtain a second diagnostic model, wherein the second training set is composed of a specific amplitude sequence, and the historical monitoring signal corresponding to the specific amplitude sequence is an abnormal working state signal of a hydro-generator; Performing a third stage of training on the second diagnostic model using a third training set, wherein the third training set consists of a target amplitude sequence, and the target amplitude sequence is a historical amplitude sequence that meets a preset condition; Among them, the preset condition is: the diagnosis result obtained by inputting the second diagnostic model is that the working state of the hydro-generator is abnormal, and then it is manually determined that the monitoring signal corresponding to the input amplitude sequence is an abnormal working state signal of the hydro-generator.

3. The data-driven method for identifying ultra-high frequency partial discharge of a hydro-turbine generator according to claim 1 is characterized in that: The types of classification networks include XGBoost, random forest, and extreme learning machine.

4. The data-driven method for identifying ultra-high frequency partial discharge of a hydro-turbine generator according to claim 1 is characterized in that: The energy characteristic is power spectral density.

5. The data-driven method for identifying ultra-high frequency partial discharge of a hydro-turbine generator according to claim 1 is characterized in that: The time domain features include at least one of phase, amplitude, waveform features, and amplitude statistical features.

6. The data-driven method for identifying ultra-high frequency partial discharge of a hydro-turbine generator according to claim 5 is characterized in that: The waveform characteristic includes one of a rise time, a fall time, a duration, and a pulse width.

7. The data-driven method for identifying ultra-high frequency partial discharge of a hydro-turbine generator according to claim 1 is characterized in that: The types of interference signals in the background noise library include corona discharge generated from the outside of the hydro-generator, electromagnetic interference, pulse interference, interference signals caused by poor grounding during monitoring signal collection, electronic noise of the collection equipment itself during monitoring signal collection, environmental noise, false signals generated by the collection equipment due to temperature changes during monitoring signal collection, false discharge signals caused by humidity changes, signal transmission interference and human interference.

8. A data-driven ultra-high frequency partial discharge identification device for a hydro-turbine generator, characterized in that: include: An acquisition module is used to obtain the monitoring signal collected by the ultra-high frequency partial discharge coupling device in the current monitoring period; A signal processing module, used for discretizing the monitoring signal to obtain an amplitude sequence; A diagnosis module, used for inputting the amplitude sequence into a pre-built signal diagnosis model to generate a diagnosis result of whether the hydro-generator is working normally; A first extraction module, used for extracting a first feature of the monitoring signal when the diagnosis result shows that the working state of the hydro-generator is abnormal; A second extraction module, configured to search for an interference signal corresponding to the first feature in the constructed background noise library, and if no interference signal corresponding to the first feature is found, extract a second feature of the monitoring signal; a signal recognition module, configured to use the second feature to match a partial discharge event corresponding to the second feature from a constructed partial discharge case library; Among them, each interference signal has its own corresponding first feature, and each partial discharge event has its own corresponding second feature. The first feature includes energy features, and the second feature includes time domain and / or frequency domain features. The signal diagnosis model is obtained by training the classification network and optimizing parameters.

9. A computer device, characterized in that: include: a memory configured to store instructions; as well as A processor is configured to call the instruction from the memory and to implement the data-driven ultra-high frequency partial discharge identification method for a hydro-turbine generator according to any one of claims 1 to 7 when executing the instruction.

10. A machine-readable storage medium, characterized in that: The machine-readable storage medium stores instructions for enabling a machine to execute the data-driven ultra-high frequency partial discharge identification method for a hydro-turbine generator according to any one of claims 1 to 7.

Images