Generator stator slot wedge tightness detection method, device and storage medium

Abstract

The present application belongs to the technical field of generators, and specifically relates to a method, device, and storage medium for detecting the tightness of a generator stator slot wedge, wherein the method comprises: S10, obtaining original sound wave data collected by a sound sensor after the slot wedge to be detected is knocked; S20, pre-processing the original sound wave data to obtain knocking sound wave data; S30, extracting target features from the knocking sound wave data to obtain sound wave feature values; wherein the target features include frequency band amplitude area, spectrum centroid, waveform factor, third peak frequency, and fifth peak frequency; S40, inputting the sound wave feature values ​​into a trained classifier to perform knocking sound wave feature classification to obtain the generator stator slot wedge tightness detection result. The method proposed in the present application is for non-destructive detection of the tightness of the generator stator slot wedge, which not only improves the detection efficiency and accuracy, but also reduces the detection cost.

Description

Technical Field

[0001] The present application belongs to the technical field of generators, and specifically relates to a method, device, and storage medium for detecting the tightness of a generator stator slot wedge. Background Art

[0002] Generators are a crucial component of power systems, primarily consisting of a stator, rotor, and other components. The stator wedges are used to secure the stator bars within the stator slots. During generator operation, the energized stator bars are subjected to radial electromagnetic forces within the transverse magnetic field within the stator slots, causing them to vibrate. Long-term vibration of the stator bars can loosen the wedges. A loosened wedge causes the stator bars to vibrate under the influence of alternating electromagnetic forces. Long-term operation of the generator damages the insulation layer, exacerbating electrical corrosion and potentially leading to breakdown of the stator bar's main insulation layer, potentially causing shutdowns and significant safety hazards.

[0003] Case studies have shown that many medium- and high-voltage motors experience magnetic wedge failure. Within three years of use, up to half of the wedges are lost. Therefore, regular inspection of the magnetic wedges is essential. Testing and re-tightening the stator wedges has become a critical component of generator maintenance. Existing inspection methods include manual tapping and measuring holes. However, with tens of thousands of stator wedges in a generator set, existing inspection methods rely heavily on operator experience, resulting in low accuracy and efficiency. Furthermore, the inspection process can cause secondary damage to the wedges, making them ineffective in meeting inspection requirements. Summary of the Invention

[0004] (1) Technical issues to be resolved

[0005] In view of the above-mentioned shortcomings and deficiencies of the prior art, the present application provides a method, device and readable storage medium for detecting the tightness of a generator stator slot wedge.

[0006] (2) Technical solution

[0007] To achieve the above objectives, this application adopts the following technical solutions:

[0008] In a first aspect, an embodiment of the present application provides a method for detecting the tightness of a stator wedge of a generator, the method comprising:

[0009] S10, obtaining original sound wave data collected by the sound sensor after the slot wedge to be detected is struck;

[0010] S20, preprocessing the original sound wave data to obtain knocking sound wave data;

[0011] S30, extracting target features from the percussion sound wave data to obtain sound wave feature values; wherein the target features include frequency band amplitude area, spectrum centroid, shape factor, third peak frequency, and fifth peak frequency;

[0012] S40: Input the acoustic wave feature value into a trained classifier to perform knocking acoustic wave feature classification, and obtain a generator stator slot wedge tightness detection result.

[0013] Optionally, S20 includes:

[0014] Using a double-threshold endpoint detection method to intercept effective knocking sound segments from the original sound wave data to obtain knocking sound wave data;

[0015] The knocking sound wave data is subjected to noise reduction processing using a second-generation wavelet transform to obtain processed knocking sound wave data.

[0016] Optionally, the classifier is a support vector machine classifier, and before S10, the step further includes training the support vector machine classifier, and the training step includes:

[0017] S01, according to the preset stator slot wedge tightness state classification, respectively detecting the stator slot wedge force in each classification state;

[0018] S02. Apply the force of the stator slot wedge in each classification state to the corrugated plate, perform a compression test on the corrugated plate, and obtain the corresponding deformation of the corrugated plate;

[0019] S03, establishing a stator slot wedge model in each classification state based on the corrugated plate deformation, and constructing a stator slot wedge tightness test platform to perform a knock test, using an excitation device to knock the stator slot wedge model to obtain a vibration sound signal in each classification state;

[0020] S04, preprocessing the vibration sound signal and extracting the vibration sound signal features to obtain a vibration signal feature value;

[0021] S05. Using the vibration signal characteristic values ​​as sample data, training a support vector machine model to obtain a trained support vector machine classifier.

[0022] Optionally, the stator slot wedge tightness status classification includes tight, slightly tight and loose; when the status is tight, the stator wire bar does not jump in both short circuit and non-short circuit conditions; when the status is slightly tight, the stator wire bar does not jump only in non-short circuit conditions; when the status is loose, the generator is prohibited from operating.

[0023] Optionally, S01 includes:

[0024] Calculate the maximum electromagnetic force F1 of the stator bar during normal operation and the maximum electromagnetic force F2 during short circuit when out-of-phase current is applied;

[0025] The electromagnetic force less than F1 is regarded as the stator wedge force when the state is loose, the electromagnetic force greater than or equal to F2 is regarded as the stator wedge force when the state is tight, and the electromagnetic force between F1 and F2 is regarded as the stator wedge force when the state is slightly tight.

[0026] Optionally, S02 includes:

[0027] A universal material testing machine is used to obtain the stress-strain characteristic curve of the corrugated plate and the relationship between the load and deformation of the corrugated plate under pressure loading.

[0028] After the pressure loading test is carried out on the testing machine, the compression curve of the test corrugated plate sample is drawn;

[0029] The actual compression curve of the corrugated plate is obtained by eliminating the deformation of the system, and S01 is calculated to obtain the average deformation of the corrugated plate in each classification state.

[0030] Optionally, extracting vibration sound signal features from the vibration sound signal in S04 includes:

[0031] Extracting time domain characteristic parameters from the vibration sound signal to characterize the tightness state of the slot wedge, the time domain characteristic parameters including root mean square value, variance, shape factor, peak factor, kurtosis, and zero crossing rate;

[0032] Extracting frequency domain characteristic parameters from the vibration sound signal to characterize the tightness of the slot wedge, the frequency domain characteristic parameters including the peak values ​​of three frequency bands and the corresponding peak frequencies, the centroid of the vertical axis spectrum, and the amplitude areas of five different frequency intervals;

[0033] Feature screening is performed based on the F-ratio and Pearson correlation coefficient method to obtain vibration sound signal features, which include frequency band amplitude area, spectrum centroid, waveform factor, third peak frequency, and fifth peak frequency.

[0034] Optionally, S05 includes:

[0035] Using the vibration signal characteristic values ​​as sample data, and dividing the sample data into a training set and a test set;

[0036] Normalize the feature parameter matrices of the training set and test set;

[0037] The training set samples were divided into three groups, and the support vector machine model parameters were optimized using the cross-validation method;

[0038] The trained support vector machine model is used to classify the test set to verify the accuracy of the trained model;

[0039] The verified support vector machine model is used as the trained support vector machine classifier.

[0040] In a second aspect, an embodiment of the present application provides a stator wedge tightness detection device, comprising: a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the generator stator wedge tightness detection method as described in any one of the first aspects above.

[0041] In a third aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the method for detecting the tightness of the stator wedge of a generator as described in any one of the first aspects above are implemented.

[0042] (3) Beneficial effects

[0043] The beneficial effects of the present application are as follows: the present application proposes a method, device and readable storage medium for detecting the tightness of the generator stator slot wedge, wherein the method comprises: S10, obtaining the original sound wave data collected by the sound sensor after the slot wedge to be detected is knocked; S20, pre-processing the original sound wave data to obtain knocking sound wave data; S30, extracting target features from the knocking sound wave data to obtain sound wave feature values; wherein the target features include frequency band amplitude area, spectrum centroid, waveform factor, third peak frequency, and fifth peak frequency; S40, inputting the sound wave feature values ​​into a trained classifier to classify the knocking sound wave features to obtain the generator stator slot wedge tightness detection result. The method proposed in the present application establishes a mechanical classification standard for the tightness of the generator stator slot wedge, and then deduces the classification standard for the corrugated plate deformation variable, thereby training a classification model. The classification standard in the present application is applicable to different generator sets, so the method of the present application can also be widely applicable to different generator sets. The method of the present application extracts the slot wedge acoustic signals under different states, and determines five characteristic parameter information after preprocessing and screening. It has fewer features and less calculation, which can improve the detection efficiency. This method performs non-destructive detection on the tightness of the generator stator slot wedge, which not only improves the accuracy but also reduces the detection cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The present application is described with the aid of the following drawings:

[0045] Figure 1 This is a flow chart of a method for detecting tightness of a generator stator slot wedge in one embodiment of the present application;

[0046] Figure 2 This is a schematic diagram of the support vector machine classifier training steps in one embodiment of the present application;

[0047] Figure 3This is a flow chart of a method for detecting tightness of a generator stator slot wedge in another embodiment of the present application;

[0048] Figure 4 This is a schematic structural diagram of a stator slot wedge model with corrugated plate fastening in another embodiment of the present application;

[0049] Figure 5 This is a graph showing actual compression curves of a corrugated plate in another embodiment of the present application;

[0050] Figure 6 This is a schematic structural diagram of a stator slot wedge tightness simulation test platform in another embodiment of the present application;

[0051] Figure 7 This is a schematic diagram of pattern recognition results in another embodiment of the present application;

[0052] Figure 8 Schematic diagram of the structure of a stator wedge tightness detection device in another embodiment of the present application.

[0053] Description of reference numerals:

[0054] 1- stator slot wedge model, 2- push-pull electromagnetic knocking device, 3- sound sensor, 4- portable dynamic acquisition system, 5- display. DETAILED DESCRIPTION

[0055] To better explain the present invention and facilitate understanding, the present invention is described in detail below through specific embodiments in conjunction with the accompanying drawings. It should be understood that the specific embodiments described below are only used to explain the relevant invention and are not intended to limit the invention. It should also be noted that the embodiments and features in the embodiments of this application can be combined with each other unless there is a conflict; for ease of description, only the parts related to the invention are shown in the drawings.

[0056] Example 1

[0057] Figure 1 FIG. 1 is a flow chart of a method for detecting the tightness of a generator stator slot wedge in one embodiment of the present application. Figure 1 As shown, the generator stator slot wedge tightness detection method of this embodiment includes:

[0058] S10, obtaining original sound wave data collected by the sound sensor after the slot wedge to be detected is struck;

[0059] S20, pre-processing the original sound wave data to obtain knocking sound wave data;

[0060] S30, extracting target features from the percussion sound wave data to obtain sound wave feature values; wherein the target features include frequency band amplitude area, spectrum centroid, waveform factor, third peak frequency, and fifth peak frequency;

[0061] S40: Input the acoustic wave feature value into a trained classifier to perform knocking acoustic wave feature classification, and obtain a generator stator slot wedge tightness detection result.

[0062] This embodiment proposes an acoustic-based method for detecting the tightness of a generator stator slot wedge, which performs non-destructive testing on the tightness of the generator stator slot wedge, thereby improving the detection efficiency and accuracy, reducing the detection cost, and achieving accurate identification of the tightness of the stator slot wedge.

[0063] In order to better understand the present invention, each step in this embodiment is described below.

[0064] In this embodiment, S20 includes:

[0065] Using a double-threshold endpoint detection method to intercept effective knocking sound segments from the original sound wave data to obtain knocking sound wave data;

[0066] The knocking sound wave data is subjected to noise reduction processing using a second-generation wavelet transform to obtain processed knocking sound wave data.

[0067] The double-threshold endpoint detection method is used to intercept effective knocking sound segments, which can reduce the amount of subsequent data calculation.

[0068] In this embodiment, the classifier is a support vector machine classifier, and before S10, the support vector machine classifier is trained. Figure 2 This is a schematic diagram of the support vector machine classifier training steps in one embodiment of the present application, as shown in FIG. Figure 2 As shown, the training steps include:

[0069] S01, according to the preset stator slot wedge tightness state classification, respectively detecting the stator slot wedge force in each classification state;

[0070] S02. Apply the force of the stator slot wedge in each classification state to the corrugated plate, perform a compression test on the corrugated plate, and obtain the corresponding deformation of the corrugated plate;

[0071] S03, establishing a stator slot wedge model in each classification state based on the corrugated plate deformation, and constructing a stator slot wedge tightness test platform to perform a knock test, using an excitation device to knock the stator slot wedge model to obtain a vibration sound signal in each classification state;

[0072] S04, preprocessing the vibration sound signal and extracting the vibration sound signal features to obtain a vibration signal feature value;

[0073] S05. Using the vibration signal characteristic values ​​as sample data, the support vector machine model is trained to obtain a trained support vector machine classifier.

[0074] The following describes the support vector machine classifier training steps in this embodiment.

[0075] In this embodiment S01, the stator slot wedge tightness status classification includes tight, slightly tight and loose; among them, when the state is tight, the stator wire bar does not jump in both short-circuit and non-short-circuit conditions; when the state is slightly tight, the stator wire bar does not jump only in non-short-circuit conditions; when the state is loose, the generator is prohibited from operating.

[0076] The method of this embodiment establishes three states of stator slot wedge tightness, which can help operators have a clearer understanding of the state of the slot wedge. Among them, the slot wedge in a slightly tight state can also be used as a key screening object in the next inspection work, providing early warning reference information for the next maintenance work.

[0077] In this embodiment, S01 may include:

[0078] Start the generator and pass out-of-phase currents through the upper and lower stator bars;

[0079] Calculate the maximum electromagnetic force F1 of the stator bar during normal operation and the maximum electromagnetic force F2 during short circuit when out-of-phase current is applied;

[0080] The electromagnetic force less than F1 is regarded as the stator wedge force when the state is loose, the electromagnetic force greater than or equal to F2 is regarded as the stator wedge force when the state is tight, and the electromagnetic force between F1 and F2 is regarded as the stator wedge force when the state is slightly tight.

[0081] In this embodiment, S02 includes:

[0082] A universal material testing machine is used to obtain the stress-strain characteristic curve of the corrugated plate and the relationship between the load and deformation of the corrugated plate under pressure loading.

[0083] After the pressure loading test is carried out on the testing machine, the compression curve of the test corrugated plate sample is drawn;

[0084] The actual compression curve of the corrugated plate is obtained by eliminating the deformation of the system, and S01 is calculated to obtain the average deformation of the corrugated plate in each classification state.

[0085] In this embodiment, the method for preprocessing the vibration sound signal may be the same as the method in S20 .

[0086] It should be noted that under test conditions, the test duration is generally longer than the time it takes for the excitation device to strike the slot wedge. Therefore, a double-threshold endpoint detection method is used to capture valid vibration sound segments, reducing the amount of subsequent data calculations. In actual application at generator maintenance sites, the environment is noisy, and the acquired percussion sound signal contains a significant amount of ambient noise, requiring noise reduction. A second-generation wavelet transform can be used for noise reduction to mitigate the effects of noise interference. For percussion sound signals acquired during laboratory simulations, noise reduction is not required.

[0087] In this embodiment, extracting the vibration sound signal features from the vibration sound signal in S04 includes:

[0088] Extracting time domain characteristic parameters from the vibration sound signal to characterize the tightness state of the slot wedge, the time domain characteristic parameters including root mean square value, variance, shape factor, peak factor, kurtosis, and zero crossing rate;

[0089] Extracting frequency domain characteristic parameters from the vibration sound signal to characterize the tightness of the slot wedge, the frequency domain characteristic parameters including the peak values ​​of three frequency bands and the corresponding peak frequencies, the centroid of the vertical axis spectrum, and the amplitude areas of five different frequency intervals;

[0090] Feature screening is performed based on the F-ratio and Pearson correlation coefficient method to obtain vibration sound signal features, which include frequency band amplitude area, spectrum centroid, waveform factor, third peak frequency, and fifth peak frequency.

[0091] In this embodiment, S05 includes:

[0092] Using the vibration signal characteristic values ​​as sample data, and dividing the sample data into a training set and a test set;

[0093] Normalize the feature parameter matrices of the training set and test set;

[0094] The training set samples were divided into three groups, and the support vector machine model parameters were optimized using the cross-validation method;

[0095] The trained support vector machine model is used to classify the test set to verify the accuracy of the trained model;

[0096] The verified support vector machine model is used as the trained support vector machine classifier.

[0097] The method of this embodiment uses SVM with optimized cross-validation parameters to classify the test set. After testing, it shows good results in accurately identifying three different slot wedge tightnesses, thereby providing an efficient and accurate detection method for complex and difficult slot wedge detection.

[0098] Example 2

[0099] This example uses slot wedge plates, corrugated plates, and gasket samples provided by China Yangtze Power Co., Ltd. to create a stator slot wedge tightness model, and conducts model training and testing, which includes the following steps:

[0100] Step 1: Establish a mechanical classification standard for the tightness and looseness of stator slot wedges.

[0101] During generator operation, the current flowing through the stator wedges can be in phase or out of phase. Analyzing and calculating these two scenarios determines the electromagnetic forces acting on the stator wedges during normal operation and short circuits, thereby establishing a mechanical classification standard for loose and tight conditions.

[0102] Step 2: Convert the mechanical classification standard into the corrugated plate deformation standard.

[0103] Based on the known mechanical state, compression tests were conducted on corrugated plate samples to obtain deformation under different force states. The mechanical classification standards were converted into corrugated plate deformation standards to facilitate the subsequent manual simulation of the stator slot wedge model under the corresponding three states.

[0104] Step 3: Experimental data collection.

[0105] An experimental platform was built, and the electromagnetic excitation device was used to strike the slot wedge plate nodes to collect the vibration and sound signals of the slot wedge plate under different states.

[0106] Step 4: Preprocessing of vibration sound signals.

[0107] In order to improve recognition accuracy, reduce noise interference and computational complexity, it is necessary to perform preprocessing such as effective segment extraction and noise reduction on the knocking sound signals obtained in the experiment.

[0108] Step 5: Feature extraction and screening.

[0109] Relevant feature parameters are extracted from the original signal to characterize the tightness or looseness information implied by the signal. The feature parameters most relevant to the evaluation index are then selected to reduce the number of feature parameters, thereby reducing the computational effort while ensuring recognition accuracy.

[0110] Step 6: Stator slot wedge tightness pattern recognition.

[0111] After feature extraction and screening, a support vector machine (SVM) is used to classify the tightness state of the sample test set, and the vibration sound signal is divided into three states and compared with the corrugated plate deformation variable standard.

[0112] Figure 3 This is a flow chart of a method for detecting the tightness of a generator stator slot wedge in another embodiment of the present application. Figure 3 Each step of this embodiment is described in detail.

[0113] Step 1: Establish a mechanical classification standard for the tightness and looseness of stator slot wedges, including:

[0114] 1) Figure 4 This is a schematic diagram of the structure of a stator slot wedge model with corrugated plate fastening in another embodiment of the present application, wherein (a) is a partial enlarged view of the gasket and the corrugated plate 12, and (b) is a schematic diagram of the overall structure of the stator slot wedge model. Figure 4 As shown, the stator slot wedge includes slot wedge 11, gasket 122, corrugated plate 121, stator bars 13, interlayer gasket 14, and iron core 15. Analysis of the electromagnetic force on the stator bars shows that when the currents flowing through the upper and lower layers of stator bars are in phase, the resultant electromagnetic force is directed toward the bottom of the slot wedge, leaving the bars unaffected.

[0115] 2) When out-of-phase current is applied, the electromagnetic force acting on the upper wire rod is directed toward the slot wedge, which is also the force acting on the corrugated plate.

[0116] 3) Further calculations show that the maximum electromagnetic force F1 = 2.0733 N / cm when the stator bars are in normal operation and the maximum electromagnetic force F2 = 248.7960 N / cm when the stator bars are in short circuit when out-of-phase current is applied.

[0117] 4) Based on F1 and F2, the slot wedge is divided into three states (≥F2, F2~F1, <F1), and a mechanical standard for classification is established.

[0118] Step 2: Establish a classification standard for slot wedge shape variables. 1) Use a universal material testing machine to obtain the stress-strain characteristic curve of the corrugated plate, and derive the relationship between the load and deformation of the corrugated plate under pressure loading, so as to compare it with the electromagnetic force exerted on the upper stator wire rod. 2) Since it is impossible to determine the ultimate pressure load when the sample corrugated plate is fully compressed, first select one of the samples to estimate the ultimate load. In addition, five samples are selected for testing. 3) After the pressure loading test on the testing machine, the compression curves of the five test corrugated plate samples are drawn, and it is found that they satisfy Hooke's law, and the system deformation also conforms to Hooke's law. 4) Figure 5 This is the actual compression curve of the corrugated plate in another embodiment of the present application. The actual compression curve of the corrugated plate is obtained by removing the deformation of the system. Figure 5 As shown, it is calculated that the average deformations corresponding to the critical forces F1 and F2 calculated in step 1 are 0.0736 mm and 1.4636 mm respectively.

[0119] Step 3: Build a test platform to conduct laboratory stator slot wedge tightness simulation test. Figure 6This is a schematic diagram of the structure of a stator slot wedge tightness simulation test platform in another embodiment of the present application. The test platform specifically includes: a stator slot wedge model 1, a push-pull electromagnetic knocking device 2, a sound sensor 3, a portable dynamic acquisition system 4, and a display 5. The electromagnetic excitation device 2 is used to knock on the slot wedge plate to generate a vibration sound signal, and the knocking sound signal is collected by the sound sensor 3. The portable dynamic acquisition system 4 stores and processes the collected signal data. The thickness of the gasket in the slot wedge is adjusted multiple times to simulate the change in the deformation of the corrugated plate, and the acoustic signals of the stator slot wedge under three different tightness conditions are collected. 100 sets of test knocks were performed for each tightness type. Therefore, a total of 300 sets of signals are generated for the three tightness conditions.

[0120] Step 4: Signal preprocessing: 1) Under the test environment, the test time is generally longer than the time it takes for the electromagnetic excitation device to knock on the slot wedge. The double-threshold endpoint detection method is used to intercept effective knocking sound segments to reduce the amount of data calculation in the later stage. 2) In order to take into account the subsequent spectrum analysis of the signal, ensure the spectrum resolution, and extract effective frequency domain features, the signal should not be intercepted too short. The effective knocking sound signal is intercepted by the double-threshold endpoint detection method, and the signal length is 0.5s. 3) During actual inspections at the generator maintenance site, the on-site environment is noisy, and the knocking sound signal obtained contains a lot of environmental noise, which requires noise reduction processing. The second-generation wavelet transform is used here for noise reduction processing to reduce the impact of noise interference. The laboratory simulation test environment has little noise and does not require noise reduction processing.

[0121] Step 5: Signal feature extraction and screening: After the sound signal is preprocessed, feature extraction and screening are required. 1) Relevant feature parameters are extracted from the original signal to characterize the loose and tight state information contained in the signal, which is mainly divided into time domain features and frequency domain features. 2) In terms of time domain features, feature parameters are extracted from six aspects: root mean square value, variance, waveform factor, peak factor, kurtosis, and zero-crossing rate. 3) In terms of spectral features, features are mainly extracted from the peaks of the three frequency bands of the signal and the corresponding peak frequencies, the centroid of the vertical axis spectrum, and the amplitude areas of five different frequency intervals. 4) After preliminary feature extraction, 18 feature signals were selected. Next, a feature screening combination method based on F-ratio and Pearson correlation coefficient method was used to evaluate and screen the extracted feature parameters. 5) Finally, after screening, unqualified feature signals were eliminated, and five feature signals were retained, namely frequency band amplitude area, spectrum centroid, waveform factor, third peak frequency, and fifth peak frequency.

[0122] Step 6 Feature pattern recognition: 1) Before model training, the feature parameter matrices of the training set and the test set are normalized. 2) The training samples are divided into 3 groups, and the parameters are optimized using the cross-validation method to obtain the feature parameters of the model. Specifically, there are 300 groups of data in three states (tight, slightly tight, and loose), 240 groups in the training set, and 60 groups in the test set. After optimization, the optimal penalty parameter C = 147.0334 and the optimal kernel function parameter g = 0.0015 are obtained. 3) The test set is classified using the SVM optimized by cross-validation optimization parameters. Figure 7 This is a schematic diagram of the results of pattern recognition in another embodiment of the present application, such as Figure 7 As shown in the figure, the support vector machine classifier is used to predict the 60 groups of data in the test set. The prediction results show that the accurate identification of three different slot wedge tightnesses shows a good effect, with an accuracy of 93.3333%.

[0123] In summary, by establishing mechanical classification standards for generator stator slot wedge tightness and corrugated plate deformation variables, we then simulated slot wedge models under different conditions for signal acquisition, preprocessing, feature extraction, and screening. Finally, using a parameter-optimized SVM method with cross-validation, we achieved excellent results in accurately identifying three different slot wedge tightnesses. This invention establishes a comprehensive solution from standard establishment to pattern recognition, which can be widely applied to a wide range of corrugated plate generator sets.

[0124] Example 3

[0125] In a second aspect, the present application provides a stator wedge tightness detection device through embodiment three, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, the steps of the generator stator wedge tightness detection method as described in any one of the above embodiments are implemented.

[0126] Figure 8 Schematic diagram of the structure of a stator wedge tightness detection device in another embodiment of the present application.

[0127] Figure 8 The stator slot wedge tightness detection device shown may include: at least one processor 101, at least one memory 102, at least one network interface 104 and another user interface 103. The various components in the electronic device are coupled together via a bus system 105. It is understood that the bus system 105 is used to achieve connection and communication between these components. In addition to including a data bus, the bus system 105 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, Figure 8 Various buses are labeled as bus system 105 .

[0128] The user interface 103 may include a display, a keyboard, or a pointing device (eg, a mouse, a trackball, or a touchpad).

[0129] It is understood that the memory 102 in this embodiment can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct RAM bus random access memory (DRRAM). The memory 62 described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0130] In some embodiments, the memory 102 stores the following elements, executable units, or data structures, or a subset thereof, or an extended set thereof: an operating system 1021 and application programs 1022 .

[0131] The operating system 1021 includes various system programs, such as a framework layer, a core library layer, and a driver layer, for implementing various basic services and handling hardware-based tasks. Application programs 1022 include various application programs for implementing various application services. Programs implementing the methods of the embodiments of the present invention may be included in application programs 1022.

[0132] In an embodiment of the present invention, the processor 101 calls a program or instruction stored in the memory 102, specifically, a program or instruction stored in the application 1022, and the processor 101 is used to execute the method steps provided in the first aspect.

[0133] The methods disclosed in the above embodiments of the present invention can be applied to or implemented by processor 101. Processor 101 may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be performed by hardware integrated logic circuits in processor 101 or by software instructions. Processor 101 may be a general-purpose processor, a digital signal processor, an application-specific integrated circuit, an off-the-shelf programmable gate array or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component. The methods, steps, and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in conjunction with the embodiments of the present invention can be directly implemented and executed by a hardware decoding processor or by a combination of hardware and software units in the decoding processor. The software units can be located in a storage medium well-known in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. This storage medium is located in memory 102. Processor 101 reads information from memory 102 and, in conjunction with its hardware, completes the steps of the above method.

[0134] In addition, in combination with the generator stator slot wedge tightness detection method in the above embodiment, an embodiment of the present invention can provide a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, any one of the generator stator slot wedge tightness detection methods in the above method embodiments is implemented.

[0135] It should be noted that in the claims, any reference signs placed between parentheses should not be construed as limiting the claims. The word "comprising" does not exclude the presence of components or steps not listed in the claim. The word "a" or "an" preceding a component does not exclude the presence of a plurality of such components. The present invention can be implemented by means of hardware comprising several distinct components and by means of a suitably programmed computer. The use of the words first, second, third, etc. is merely for convenience and does not imply any order. These words should be understood as part of the component name.

[0136] In addition, it should be noted that, in the description of this specification, the description of the terms "one embodiment", "some embodiments", "embodiment", "example", "specific example" or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and the features of different embodiments or examples, unless they are contradictory.

[0137] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments after learning the basic creative concept. Therefore, the claims should be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0138] Obviously, those skilled in the art may make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if such modifications and variations fall within the scope of the claims and their equivalents, the present invention shall also include such modifications and variations.

Claims

1. A method for detecting the tightness of a generator stator slot wedge, characterized in that: The method includes: S10, obtaining original sound wave data collected by the sound sensor after the slot wedge to be detected is struck; S20, preprocessing the original sound wave data to obtain knocking sound wave data; S30, extracting target features from the percussion sound wave data to obtain sound wave feature values; wherein the target features include frequency band amplitude area, spectrum centroid, shape factor, third peak frequency, and fifth peak frequency; S40, inputting the acoustic wave feature value into a trained classifier to perform knocking acoustic wave feature classification, and obtaining a generator stator slot wedge tightness detection result; The classifier is a support vector machine classifier. Before S10, the support vector machine classifier is trained. The training steps include: S01. According to a preset stator slot wedge looseness and tightness classification standard, the force on the stator slot wedge in each classification state is detected. During the operation of the generator, the current flowing into the stator slot wedge internal bars may be in phase or out of phase. These two situations are analyzed and calculated to determine the electromagnetic force on the bars during normal operation and short circuit of the generator, thereby establishing a mechanical classification standard for looseness and tightness. S02. Applying the force of the stator slot wedge under each classification state to the corrugated plate to perform a compression test on the corrugated plate to obtain the corresponding corrugated plate deformation; specifically, performing compression tests on the corrugated plate samples to obtain deformations under different force states, and converting the mechanical classification standard into a corrugated plate deformation standard; S03, establishing a stator slot wedge model in each classification state based on the corrugated plate deformation, and constructing a stator slot wedge tightness test platform to perform a knock test, using an excitation device to knock the stator slot wedge model to obtain a vibration sound signal in each classification state; S04, preprocessing the vibration sound signal and extracting the vibration sound signal features to obtain a vibration signal feature value; S05. Using the vibration signal characteristic values ​​as sample data, training a support vector machine model to obtain a trained support vector machine classifier.

2. The method for detecting tightness of a generator stator slot wedge according to claim 1, characterized in that: The S20 includes: Using a double-threshold endpoint detection method to intercept effective knocking sound segments from the original sound wave data to obtain knocking sound wave data; The knocking sound wave data is subjected to noise reduction processing using a second-generation wavelet transform to obtain processed knocking sound wave data.

3. The method for detecting tightness of a generator stator wedge according to claim 1, characterized in that: The stator slot wedge tightness status classification includes tight, slightly tight and loose; among them, when the state is tight, the stator wire bar does not jump in short-circuit and non-short-circuit conditions; when the state is slightly tight, the stator wire bar does not jump only in non-short-circuit conditions; when the state is loose, the generator is prohibited from operating.

4. The method for detecting tightness of a generator stator slot wedge according to claim 3, characterized in that: S01 includes: Calculate the maximum electromagnetic force F1 of the stator bar during normal operation and the maximum electromagnetic force F2 during short circuit when out-of-phase current is applied; The electromagnetic force less than F1 is regarded as the stator wedge force when the state is loose, the electromagnetic force greater than or equal to F2 is regarded as the stator wedge force when the state is tight, and the electromagnetic force between F1 and F2 is regarded as the stator wedge force when the state is slightly tight.

5. The method for detecting tightness of a generator stator slot wedge according to claim 1, characterized in that: S02 includes: A universal material testing machine is used to obtain the stress-strain characteristic curve of the corrugated plate and the relationship between the load and deformation of the corrugated plate under pressure loading. After the pressure loading test is carried out on the testing machine, the compression curve of the test corrugated plate sample is drawn; The actual compression curve of the corrugated plate is obtained by eliminating the deformation of the system, and S01 is calculated to obtain the average deformation of the corrugated plate in each classification state.

6. The method for detecting tightness of a generator stator slot wedge according to claim 1, characterized in that: Extracting the vibration sound signal features from the vibration sound signal in S04 includes: Extracting time domain characteristic parameters from the vibration sound signal to characterize the tightness state of the slot wedge, the time domain characteristic parameters including root mean square value, variance, shape factor, peak factor, kurtosis, and zero crossing rate; Extracting frequency domain characteristic parameters from the vibration sound signal to characterize the tightness of the slot wedge, the frequency domain characteristic parameters including the peak values ​​of three frequency bands and the corresponding peak frequencies, the centroid of the vertical axis spectrum, and the amplitude areas of five different frequency intervals; Feature screening is performed based on the F-ratio and Pearson correlation coefficient method to obtain vibration sound signal features, which include frequency band amplitude area, spectrum centroid, waveform factor, third peak frequency, and fifth peak frequency.

7. The method for detecting tightness of a generator stator slot wedge according to claim 1, characterized in that: S05 includes: Using the vibration signal characteristic values ​​as sample data, and dividing the sample data into a training set and a test set; Normalize the feature parameter matrices of the training set and test set; The training set samples were divided into three groups, and the support vector machine model parameters were optimized using the cross-validation method; The trained support vector machine model is used to classify the test set to verify the accuracy of the trained model; The verified support vector machine model is used as the trained support vector machine classifier.

8. A stator slot wedge tightness detection device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the steps of the method for detecting the tightness of a generator stator slot wedge are implemented as described in any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for detecting the tightness of a generator stator wedge are implemented as described in any one of claims 1 to 7.

Images