Methods, apparatus, equipment, storage media, and program products for optimizing the detection threshold of partial discharge in generator stator.

Abstract

This application relates to a method, apparatus, device, storage medium, and program product for optimizing the detection threshold of partial discharge in generator stator, belonging to the field of power equipment condition monitoring technology. This method can improve the dynamic adaptability of threshold adjustment. The method includes: preprocessing the original partial discharge signal collected by the partial discharge sensor of the generator under test to obtain a target partial discharge signal; performing feature extraction and pattern recognition on the target partial discharge signal to obtain the statistical characteristics, frequency domain characteristics, and discharge type of the generator stator partial discharge; calculating the false alarm rate and false negative rate of the generator stator partial discharge using a stochastic gradient descent algorithm based on the statistical characteristics, frequency domain characteristics, and discharge type, and determining a comprehensive loss function based on the false alarm rate and false negative rate; optimizing the initial detection threshold based on the learning rate of the stochastic gradient descent algorithm and the comprehensive loss function to obtain an updated target detection threshold.

Description

Technical Field

[0001] This application relates to the field of power equipment condition monitoring technology, and in particular to an optimization method, apparatus, computer equipment, computer-readable storage medium, and computer program product for the detection threshold of partial discharge in generator stator. Background Technology

[0002] Partial discharge (PD) in generator stator refers to the partial discharge phenomenon occurring inside the stator windings or between the windings and the stator core. This phenomenon is usually caused by the deterioration or aging of electrical insulation materials and is a common type of generator insulation fault. Partial discharge in generator stator is an important sign of insulation deterioration, and its detection is crucial for preventing equipment failure.

[0003] Traditional detection methods mainly use static models to set and adjust the partial discharge detection threshold. However, this method relies on offline training data, which makes it difficult to cope with real-time operating condition fluctuations. Furthermore, model updates require manual intervention, resulting in poor dynamic adaptability. Summary of the Invention

[0004] Therefore, it is necessary to provide an optimization method, apparatus, computer equipment, computer-readable storage medium, and computer program product for the partial discharge detection threshold of generator stator in response to the above-mentioned technical problems.

[0005] In a first aspect, this application provides a method for optimizing the detection threshold of partial discharge in a generator stator, including:

[0006] The original partial discharge signal collected by the partial discharge sensor of the generator under test is acquired, and the original partial discharge signal is preprocessed to obtain the target partial discharge signal.

[0007] Feature extraction and pattern recognition are performed on the target partial discharge signal to obtain the statistical characteristics, frequency domain characteristics, and discharge type of generator stator partial discharge of the target partial discharge signal;

[0008] Based on the statistical characteristics, the frequency domain characteristics, and the discharge type, the false alarm rate and the missed alarm rate of the generator stator partial discharge are calculated using the stochastic gradient descent algorithm, and the comprehensive loss function is determined based on the false alarm rate and the missed alarm rate.

[0009] An initial detection threshold is obtained, and the initial detection threshold is optimized based on the learning rate of the stochastic gradient descent algorithm and the comprehensive loss function to obtain an updated target detection threshold.

[0010] In one embodiment, optimizing the initial detection threshold based on the learning rate of the stochastic gradient descent algorithm and the comprehensive loss function to obtain an updated target detection threshold includes:

[0011] The initial detection threshold, the learning rate of the stochastic gradient descent algorithm, and the comprehensive loss function are input into a preset threshold optimization model for threshold optimization calculation to obtain the target threshold value output by the threshold optimization model; a threshold update configuration file is generated based on the target threshold value; in response to the threshold update instruction sent by the server, the initial detection threshold is updated according to the threshold update configuration file to obtain the updated target detection threshold.

[0012] In one embodiment, the method further comprises:

[0013] The current partial discharge monitoring value of the generator under test is obtained, and the relationship between the current partial discharge monitoring value and the target detection threshold is determined. If the current partial discharge monitoring value exceeds the target detection threshold, a real-time alarm signal and real-time alarm information are generated.

[0014] In one embodiment, obtaining the initial detection threshold includes:

[0015] The historical partial discharge monitoring values ​​and online partial discharge monitoring values ​​of the generator under test are obtained. Training samples are generated based on the historical partial discharge monitoring values ​​and the online partial discharge monitoring values. An initial threshold configuration file is obtained by training based on the training samples, and the initial detection threshold is obtained based on the initial threshold configuration file.

[0016] In one embodiment, the step of preprocessing the original partial discharge signal to obtain the target partial discharge signal includes:

[0017] According to the preset cutoff frequency, the original partial discharge signal is subjected to anti-aliasing filtering to obtain the filtered partial discharge signal; the filtered partial discharge signal is amplified according to the target gain, and the amplified partial discharge signal is subjected to analog-to-digital conversion to obtain the target partial discharge signal.

[0018] In one embodiment, the method further comprises:

[0019] Based on the false alarm rate and the false alarm rate of the partial discharge of the generator stator, a feedback signal is generated, and based on the feedback signal, the stochastic gradient descent algorithm is driven to update the parameters, so as to obtain the updated stochastic gradient descent algorithm.

[0020] The step of calculating the false alarm rate and false negative rate of partial discharge of the generator stator using a stochastic gradient descent algorithm based on the statistical characteristics, frequency domain characteristics, and discharge type includes:

[0021] The statistical features, the frequency domain features, and the discharge type are input into the updated stochastic gradient descent algorithm to obtain the false alarm rate and the missed alarm rate of the generator stator partial discharge.

[0022] Secondly, this application also provides an optimization device for the detection threshold of partial discharge in a generator stator, comprising:

[0023] The signal processing module is used to acquire the original partial discharge signal collected by the partial discharge sensor of the generator under test, and to perform signal preprocessing on the original partial discharge signal to obtain the target partial discharge signal;

[0024] The feature extraction module is used to extract features and recognize patterns from the target partial discharge signal to obtain the statistical features, frequency domain features, and discharge type of the generator stator partial discharge of the target partial discharge signal.

[0025] The function determination module is used to calculate the false alarm rate and false negative rate of partial discharge of the generator stator using a stochastic gradient descent algorithm based on the statistical characteristics, the frequency domain characteristics and the discharge type, and to determine the comprehensive loss function based on the false alarm rate and the false negative rate.

[0026] The threshold optimization module is used to obtain an initial detection threshold, and optimize the initial detection threshold according to the learning rate of the stochastic gradient descent algorithm and the comprehensive loss function to obtain an updated target detection threshold.

[0027] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0028] The original partial discharge signal collected by the partial discharge sensor of the generator under test is acquired, and the original partial discharge signal is preprocessed to obtain the target partial discharge signal.

[0029] Feature extraction and pattern recognition are performed on the target partial discharge signal to obtain the statistical characteristics, frequency domain characteristics, and discharge type of generator stator partial discharge of the target partial discharge signal;

[0030] Based on the statistical characteristics, the frequency domain characteristics, and the discharge type, the false alarm rate and the missed alarm rate of the generator stator partial discharge are calculated using the stochastic gradient descent algorithm, and the comprehensive loss function is determined based on the false alarm rate and the missed alarm rate.

[0031] An initial detection threshold is obtained, and the initial detection threshold is optimized based on the learning rate of the stochastic gradient descent algorithm and the comprehensive loss function to obtain an updated target detection threshold.

[0032] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0033] The original partial discharge signal collected by the partial discharge sensor of the generator under test is acquired, and the original partial discharge signal is preprocessed to obtain the target partial discharge signal.

[0034] Feature extraction and pattern recognition are performed on the target partial discharge signal to obtain the statistical characteristics, frequency domain characteristics, and discharge type of generator stator partial discharge of the target partial discharge signal;

[0035] Based on the statistical characteristics, the frequency domain characteristics, and the discharge type, the false alarm rate and the missed alarm rate of the generator stator partial discharge are calculated using the stochastic gradient descent algorithm, and the comprehensive loss function is determined based on the false alarm rate and the missed alarm rate.

[0036] An initial detection threshold is obtained, and the initial detection threshold is optimized based on the learning rate of the stochastic gradient descent algorithm and the comprehensive loss function to obtain an updated target detection threshold.

[0037] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0038] The original partial discharge signal collected by the partial discharge sensor of the generator under test is acquired, and the original partial discharge signal is preprocessed to obtain the target partial discharge signal.

[0039] Feature extraction and pattern recognition are performed on the target partial discharge signal to obtain the statistical characteristics, frequency domain characteristics, and discharge type of generator stator partial discharge of the target partial discharge signal;

[0040] Based on the statistical characteristics, the frequency domain characteristics, and the discharge type, the false alarm rate and the missed alarm rate of the generator stator partial discharge are calculated using the stochastic gradient descent algorithm, and the comprehensive loss function is determined based on the false alarm rate and the missed alarm rate.

[0041] An initial detection threshold is obtained, and the initial detection threshold is optimized based on the learning rate of the stochastic gradient descent algorithm and the comprehensive loss function to obtain an updated target detection threshold.

[0042] The aforementioned method, apparatus, computer equipment, computer-readable storage medium, and computer program product for optimizing the generator stator partial discharge detection threshold obtains the statistical characteristics, frequency domain characteristics, and discharge type of the target partial discharge signal by performing feature extraction and pattern recognition on the target partial discharge signal. Then, based on the statistical characteristics, frequency domain characteristics, and discharge type, the false alarm rate and false negative rate of the generator stator partial discharge are calculated using a stochastic gradient descent algorithm. A comprehensive loss function is determined based on the false alarm rate and false negative rate. An initial detection threshold is then obtained, and the initial detection threshold is optimized based on the learning rate of the stochastic gradient descent algorithm and the comprehensive loss function to obtain an updated target detection threshold. This application achieves adaptive optimization and updating of the partial discharge detection threshold through a closed-loop feedback mechanism and a stochastic gradient descent algorithm. It can automatically adjust the generator stator partial discharge detection threshold, cope with real-time changes such as equipment aging and load fluctuations, and significantly shorten the adjustment cycle of the detection threshold, thereby improving the dynamic adaptability of the threshold adjustment and reducing the false alarm rate and false negative rate of generator stator partial discharge. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 This is an application environment diagram of an optimization method for the generator stator partial discharge detection threshold in one embodiment;

[0045] Figure 2 This is a flowchart illustrating an optimization method for the generator stator partial discharge detection threshold in one embodiment;

[0046] Figure 3 This is a flowchart illustrating the threshold optimization steps in one embodiment;

[0047] Figure 4 This is a flowchart illustrating an optimization method for the generator stator partial discharge detection threshold in a specific embodiment.

[0048] Figure 5 This is a flowchart illustrating an optimization method for the partial discharge detection threshold of a generator stator in an application embodiment.

[0049] Figure 6 This is a structural block diagram of an optimization device for the generator stator partial discharge detection threshold in one embodiment;

[0050] Figure 7 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0052] The optimization method for generator stator partial discharge detection threshold provided in this application embodiment can be applied to, for example, Figure 1 The application environment shown depicts a scenario where the terminal communicates with the server via a network. The data storage system stores the data the server needs to process. This data storage system can be integrated onto the server or hosted on the cloud or other network servers. In such an environment... Figure 1 In the application environment shown, the terminal can be, but is not limited to, various personal computers, laptops, smartphones, and tablets. The server can be implemented using a standalone server or a server cluster consisting of multiple servers.

[0053] In one embodiment, such as Figure 2 As shown, an optimization method for the partial discharge detection threshold of a generator stator is provided, which is then applied to... Figure 1 Taking the terminal in the example, the explanation includes the following steps:

[0054] Step S201: Obtain the original partial discharge signal collected by the partial discharge sensor of the generator under test, perform signal preprocessing on the original partial discharge signal to obtain the target partial discharge signal.

[0055] Among them, the partial discharge sensor can be a high-frequency pulse current sensor built into the generator stator winding, and its sampling rate can be set to 40MHz-350MHz.

[0056] Specifically, the terminal is set to collect a partial discharge signal every 50ms when the generator under test is running, and to acquire the original partial discharge signal collected by the high-frequency pulse current sensor. The original partial discharge signal is then preprocessed by signal filtering, signal amplification, and analog-to-digital conversion to obtain the target partial discharge signal.

[0057] Step S202: Feature extraction and pattern recognition are performed on the target partial discharge signal to obtain the statistical characteristics, frequency domain characteristics, and discharge type of generator stator partial discharge of the target partial discharge signal.

[0058] Among them, the statistical characteristics include, but are not limited to, the peak value and pulse repetition rate of the discharge pulse of the target partial discharge signal, and the frequency domain characteristics can be the spectral energy distribution of the target partial discharge signal in the range of 50-1000kHz.

[0059] Specifically, the terminal analyzes the statistical and frequency domain characteristics of the target partial discharge signal to obtain the discharge pulse peak value and pulse repetition rate of the target partial discharge signal, and classifies the discharge type of the target partial discharge signal based on wavelet packet decomposition to obtain the discharge type corresponding to the generator stator partial discharge.

[0060] Step S203: Based on statistical characteristics, frequency domain characteristics, and discharge type, calculate the false alarm rate and false negative rate of partial discharge of generator stator using the stochastic gradient descent algorithm, and determine the comprehensive loss function based on the false alarm rate and false negative rate.

[0061] The overall loss function can be expressed as (0.7FPR + 0.3FNR), where FPR is the false alarm rate and FNR is the false negative rate.

[0062] Specifically, the terminal uses an improved stochastic gradient descent algorithm to calculate the false alarm rate and missed alarm rate of generator stator partial discharge based on statistical characteristics, frequency domain characteristics and discharge type, and determines the comprehensive loss function based on the false alarm rate and missed alarm rate.

[0063] Step S204: Obtain the initial detection threshold. Optimize the initial detection threshold based on the learning rate of the stochastic gradient descent algorithm and the comprehensive loss function to obtain the updated target detection threshold.

[0064] Among them, stochastic gradient descent (SGD) is an efficient optimization algorithm that can be applied to machine learning and deep learning. The learning rate of stochastic gradient descent is a learning rate that decays over time and is used to control the step size of parameter updates.

[0065] Specifically, the terminal dynamically generates training samples by combining online monitoring data and historical databases, and uses the training samples for training to generate an initial detection threshold. Then, based on the learning rate of the stochastic gradient descent algorithm and the comprehensive loss function, the initial detection threshold is optimized through a preset threshold optimization model to obtain the updated target detection threshold.

[0066] In the aforementioned optimization method for the generator stator partial discharge detection threshold, feature extraction and pattern recognition are performed on the target partial discharge signal to obtain its statistical characteristics, frequency domain characteristics, and discharge type. Then, based on the statistical characteristics, frequency domain characteristics, and discharge type, the false alarm rate and false negative rate of the generator stator partial discharge are calculated using a stochastic gradient descent algorithm. A comprehensive loss function is then determined based on these rates. Finally, an initial detection threshold is obtained, and it is optimized using the learning rate of the stochastic gradient descent algorithm and the comprehensive loss function to obtain an updated target detection threshold. This application achieves adaptive optimization of the partial discharge detection threshold through a closed-loop feedback mechanism and a stochastic gradient descent algorithm. This allows for automatic adjustment of the generator stator partial discharge detection threshold, enabling it to cope with real-time changes such as equipment aging and load fluctuations. Furthermore, it significantly shortens the threshold adjustment cycle, thereby improving the dynamic adaptability of the threshold adjustment and reducing the false alarm rate and false negative rate of generator stator partial discharge.

[0067] In one embodiment, Figure 3 As shown, in step S204 above, the initial detection threshold is optimized based on the learning rate of the stochastic gradient descent algorithm and the comprehensive loss function to obtain the updated target detection threshold. This specifically includes the following steps:

[0068] Step S301: Input the initial detection threshold, the learning rate of the stochastic gradient descent algorithm, and the comprehensive loss function into the preset threshold optimization model to perform threshold optimization calculation, and obtain the threshold target value output by the threshold optimization model.

[0069] Step S302: Generate a threshold update configuration file based on the threshold target value.

[0070] Step S303: In response to the threshold update instruction sent by the server, the initial detection threshold is updated according to the threshold update configuration file to obtain the updated target detection threshold.

[0071] The preset threshold optimization model can be... In the above formula The initial detection threshold is given in pC (picocoulombs). Let η(t) be the threshold target value (in pC), η(t) be the learning rate that decays over time, and J be the comprehensive loss function (0.7FPR + 0.3FNR).

[0072] Specifically, the terminal determines the learning rate of the stochastic gradient descent algorithm, inputs the initial detection threshold, the learning rate of the stochastic gradient descent algorithm, and the comprehensive loss function into a preset threshold optimization model to perform threshold optimization calculation, and obtains the threshold target value output by the threshold optimization model; based on the threshold target value, a threshold update configuration file in JSON format is generated; in response to the threshold update instruction sent by the server, the initial detection threshold is updated according to the threshold update configuration file to obtain the updated target detection threshold.

[0073] In one embodiment, the method of this application further includes:

[0074] The system acquires the current partial discharge monitoring value of the generator under test and determines the relationship between the current partial discharge monitoring value and the target detection threshold. If the current partial discharge monitoring value exceeds the target detection threshold, a real-time alarm signal and real-time alarm information are generated.

[0075] The real-time alarm signal can be presented as a light alarm or an audible alarm.

[0076] Specifically, the terminal acquires the current partial discharge monitoring value of the generator under test in real time and determines the relationship between the current partial discharge monitoring value and the target detection threshold; when it is detected that the current partial discharge monitoring value exceeds the target detection threshold, it generates a real-time alarm signal and real-time alarm information.

[0077] In one embodiment, step S204 above, obtaining the initial detection threshold, specifically includes the following steps:

[0078] The historical and online partial discharge monitoring values ​​of the generator under test are obtained. Training samples are generated based on the historical and online partial discharge monitoring values. An initial threshold configuration file is obtained by training based on the training samples, and an initial detection threshold is obtained based on the initial threshold configuration file.

[0079] Specifically, the terminal acquires historical partial discharge monitoring values ​​and online partial discharge monitoring values ​​of the generator under test, and dynamically generates training samples based on the historical partial discharge monitoring values ​​and online partial discharge monitoring values; it then trains based on the training samples to obtain an initial threshold configuration file, and generates an initial detection threshold based on the initial threshold configuration file.

[0080] In one embodiment, step S201 above involves preprocessing the original partial discharge signal to obtain the target partial discharge signal, specifically including the following steps:

[0081] Based on the preset cutoff frequency, the original partial discharge signal is subjected to anti-aliasing filtering to obtain the filtered partial discharge signal; based on the target gain, the filtered partial discharge signal is amplified, and the amplified partial discharge signal is subjected to analog-to-digital conversion to obtain the target partial discharge signal.

[0082] The preset cutoff frequency can be set to 4MHz, the signal amplification gain can be set to 20dB, and analog-to-digital conversion can be performed using a 16-bit ADC.

[0083] Specifically, the terminal performs anti-aliasing filtering on the original partial discharge signal according to a preset cutoff frequency to obtain a filtered partial discharge signal; then, it amplifies the filtered partial discharge signal according to the target gain, and performs analog-to-digital conversion on the amplified partial discharge signal to obtain the target partial discharge signal.

[0084] In one embodiment, the method of this application further includes:

[0085] Based on the false alarm rate and false alarm rate of partial discharge of generator stator, a feedback signal is generated, and based on the feedback signal, the stochastic gradient descent algorithm is driven to update the parameters, resulting in the updated stochastic gradient descent algorithm.

[0086] In step S203 above, based on statistical characteristics, frequency domain characteristics, and discharge type, the false alarm rate and false negative rate of generator stator partial discharge are calculated using the stochastic gradient descent algorithm. Specifically, this includes the following steps:

[0087] By inputting statistical characteristics, frequency domain characteristics, and discharge type into the updated stochastic gradient descent algorithm, the false alarm rate and false negative rate of partial discharge of generator stator are obtained.

[0088] Specifically, the terminal generates false alarm / missed alarm results based on the false alarm rate and missed alarm rate of generator stator partial discharge, uses the false alarm / missed alarm results as feedback signals, and drives the stochastic gradient descent algorithm to update parameters based on the feedback signals, thus obtaining the updated stochastic gradient descent algorithm; inputting statistical features, frequency domain features, and discharge type into the updated stochastic gradient descent algorithm, the false alarm rate and missed alarm rate of generator stator partial discharge are obtained.

[0089] In one embodiment, such as Figure 4 As shown, an optimization method for the generator stator partial discharge detection threshold is provided in a specific embodiment, which includes the following steps:

[0090] Step S401: Obtain the original partial discharge signal collected by the partial discharge sensor of the generator under test; perform anti-aliasing filtering on the original partial discharge signal according to the preset cutoff frequency to obtain the filtered partial discharge signal; perform signal amplification on the filtered partial discharge signal according to the target gain; and perform analog-to-digital conversion on the amplified partial discharge signal to obtain the target partial discharge signal.

[0091] Step S402: Perform feature extraction and pattern recognition on the target partial discharge signal to obtain the statistical characteristics, frequency domain characteristics, and discharge type of generator stator partial discharge of the target partial discharge signal.

[0092] Step S403: Based on statistical characteristics, frequency domain characteristics, and discharge type, calculate the false alarm rate and false negative rate of partial discharge of generator stator using the stochastic gradient descent algorithm, and determine the comprehensive loss function based on the false alarm rate and false negative rate.

[0093] Step S404: Obtain historical partial discharge monitoring values ​​and online partial discharge monitoring values ​​of the generator under test; generate training samples based on historical partial discharge monitoring values ​​and online partial discharge monitoring values; train based on training samples to obtain an initial threshold configuration file; and obtain an initial detection threshold based on the initial threshold configuration file.

[0094] Step S405: Input the initial detection threshold, the learning rate of the stochastic gradient descent algorithm, and the comprehensive loss function into the preset threshold optimization model to perform threshold optimization calculation, and obtain the threshold target value output by the threshold optimization model; generate a threshold update configuration file based on the threshold target value; respond to the threshold update instruction sent by the server, update the initial detection threshold according to the threshold update configuration file to obtain the updated target detection threshold.

[0095] Step S406: Obtain the current partial discharge monitoring value of the generator under test, and determine the relationship between the current partial discharge monitoring value and the target detection threshold; if the current partial discharge monitoring value exceeds the target detection threshold, generate a real-time alarm signal and real-time alarm information.

[0096] The beneficial effects of the above embodiments are as follows:

[0097] This application achieves adaptive optimization and updating of the partial discharge detection threshold through a closed-loop feedback mechanism and a stochastic gradient descent algorithm. It can automatically adjust the detection threshold of generator stator partial discharge, cope with real-time changes such as equipment aging and load fluctuations, and greatly shorten the adjustment cycle of the detection threshold, thereby improving the dynamic adaptability of threshold adjustment and reducing the false alarm rate and false negative rate of generator stator partial discharge.

[0098] To more clearly illustrate the optimization method for generator stator partial discharge detection threshold provided in this application, a specific embodiment is used below to describe the optimization method for generator stator partial discharge detection threshold. As an application embodiment, such as... Figure 5 As shown, another method for optimizing the generator stator partial discharge detection threshold is provided, which specifically includes the following steps:

[0099] Step 1, Initialization Phase (T0):

[0100] (1) Load historical partial discharge data on the server to train the initial threshold (default threshold 100pC).

[0101] (2) Edge computing terminal obtains initial threshold configuration.

[0102] Step 2, Real-time Execution Phase (T1-Tn):

[0103] Execution sequence:

[0104] Edge computing terminal -> server: Send feature vector (every 10 seconds);

[0105] Server-side -> Server-side: Online learning computing;

[0106] Server-side -> Edge computing terminal: Returns the updated target detection threshold;

[0107] Edge computing terminal -> Actuator: Update detection threshold;

[0108] Actuator -> Sensor: Continue monitoring.

[0109] Step 3, Feedback Optimization Cycle:

[0110] Data acquisition frequency: 10MHz (configurable);

[0111] Threshold update frequency: 1Hz (configurable).

[0112] The beneficial effects of the above embodiments are as follows:

[0113] 1) Improved dynamic adaptability: The threshold adjustment cycle has been shortened from "days" to "minutes", reducing the false alarm rate by 30%.

[0114] 2) Reduced operation and maintenance costs: Eliminating manual intervention and self-evolving models save 90% of maintenance time.

[0115] 3) Enhanced long-term stability: By resisting data distribution shift through online learning, the model accuracy decay rate is reduced to 2% / year.

[0116] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0117] Based on the same inventive concept, this application also provides an optimization device for the generator stator partial discharge detection threshold, which is used to implement the optimization method for the generator stator partial discharge detection threshold described above. The solution provided by this device is similar to the implementation described in the above method. Therefore, the specific limitations in one or more embodiments of the generator stator partial discharge detection threshold optimization device provided below can be found in the limitations of the generator stator partial discharge detection threshold optimization method described above, and will not be repeated here.

[0118] In one exemplary embodiment, such as Figure 6 As shown, an optimization device for the detection threshold of partial discharge in generator stator is provided, comprising:

[0119] The signal processing module 601 is used to acquire the original partial discharge signal collected by the partial discharge sensor of the generator under test, perform signal preprocessing on the original partial discharge signal, and obtain the target partial discharge signal.

[0120] The feature extraction module 602 is used to extract features and recognize patterns from the target partial discharge signal to obtain the statistical features, frequency domain features, and discharge type of the generator stator partial discharge of the target partial discharge signal.

[0121] The function determination module 603 is used to calculate the false alarm rate and false negative rate of partial discharge of generator stator through stochastic gradient descent algorithm based on statistical characteristics, frequency domain characteristics and discharge type, and to determine the comprehensive loss function based on the false alarm rate and false negative rate.

[0122] The threshold optimization module 604 is used to obtain the initial detection threshold, optimize the initial detection threshold according to the learning rate of the stochastic gradient descent algorithm and the comprehensive loss function, and obtain the updated target detection threshold.

[0123] In one embodiment, the threshold optimization module 604 is further configured to input the initial detection threshold, the learning rate of the stochastic gradient descent algorithm, and the comprehensive loss function into a preset threshold optimization model to perform threshold optimization calculation, thereby obtaining the threshold target value output by the threshold optimization model; generate a threshold update configuration file based on the threshold target value; and update the initial detection threshold according to the threshold update configuration file in response to the threshold update instruction sent by the server, thereby obtaining the updated target detection threshold.

[0124] In one embodiment, the device for optimizing the generator stator partial discharge detection threshold further includes a real-time alarm module, which is used to acquire the current partial discharge monitoring value of the generator under test, determine the relationship between the current partial discharge monitoring value and the target detection threshold, and generate a real-time alarm signal and real-time alarm information when it is detected that the current partial discharge monitoring value exceeds the target detection threshold.

[0125] In one embodiment, the threshold optimization module 604 is further configured to acquire historical partial discharge monitoring values ​​and online partial discharge monitoring values ​​of the generator under test, generate training samples based on the historical partial discharge monitoring values ​​and online partial discharge monitoring values, train based on the training samples to obtain an initial threshold configuration file, and obtain an initial detection threshold based on the initial threshold configuration file.

[0126] In one embodiment, the signal processing module 601 is further configured to perform anti-aliasing filtering on the original partial discharge signal according to a preset cutoff frequency to obtain a filtered partial discharge signal; perform signal amplification on the filtered partial discharge signal according to a target gain; and perform analog-to-digital conversion on the amplified partial discharge signal to obtain a target partial discharge signal.

[0127] In one embodiment, the optimization device for the generator stator partial discharge detection threshold further includes an algorithm update module, which generates a feedback signal based on the false alarm rate and false negative rate of the generator stator partial discharge, and drives the stochastic gradient descent algorithm to update its parameters based on the feedback signal, thereby obtaining the updated stochastic gradient descent algorithm; the function determination module 603 is further used to input statistical features, frequency domain features and discharge type into the updated stochastic gradient descent algorithm to obtain the false alarm rate and false negative rate of the generator stator partial discharge.

[0128] Each module in the aforementioned optimization device for generator stator partial discharge detection threshold can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0129] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 7 As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When executed by the processor, the computer program implements an optimization method for the detection threshold of partial discharge in a generator stator. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device casing, or an external keyboard, touchpad or mouse.

[0130] Those skilled in the art will understand that Figure 7 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0131] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0132] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0133] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0134] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0135] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0136] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0137] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. An optimization method for the detection threshold of partial discharge in generator stator, characterized in that, The method comprises: The original partial discharge signal collected by the partial discharge sensor of the generator under test is acquired, and the original partial discharge signal is preprocessed to obtain the target partial discharge signal. Feature extraction and pattern recognition are performed on the target partial discharge signal to obtain the statistical characteristics, frequency domain characteristics, and discharge type of generator stator partial discharge of the target partial discharge signal; Based on the statistical characteristics, the frequency domain characteristics, and the discharge type, the false alarm rate and the missed alarm rate of the generator stator partial discharge are calculated using the stochastic gradient descent algorithm, and the comprehensive loss function is determined based on the false alarm rate and the missed alarm rate. An initial detection threshold is obtained, and the initial detection threshold is optimized based on the learning rate of the stochastic gradient descent algorithm and the comprehensive loss function to obtain an updated target detection threshold.

2. The method according to claim 1, characterized in that, The step of optimizing the initial detection threshold based on the learning rate of the stochastic gradient descent algorithm and the comprehensive loss function to obtain the updated target detection threshold includes: The initial detection threshold, the learning rate of the stochastic gradient descent algorithm, and the comprehensive loss function are input into a preset threshold optimization model to perform threshold optimization calculation, thereby obtaining the threshold target value output by the threshold optimization model. Generate a threshold update configuration file based on the target threshold value; In response to the threshold update instruction sent by the server, the initial detection threshold is updated according to the threshold update configuration file to obtain the updated target detection threshold.

3. The method according to claim 2, characterized in that, The method further includes: Obtain the current partial discharge monitoring value of the generator under test, and determine the relationship between the current partial discharge monitoring value and the target detection threshold; If the current partial discharge monitoring value is detected to exceed the target detection threshold, a real-time alarm signal and real-time alarm information are generated.

4. The method according to claim 1, characterized in that, The process of obtaining the initial detection threshold includes: The historical partial discharge monitoring values ​​and online partial discharge monitoring values ​​of the generator under test are obtained, and training samples are generated based on the historical partial discharge monitoring values ​​and the online partial discharge monitoring values. The initial threshold configuration file is obtained by training based on the training samples, and the initial detection threshold is obtained according to the initial threshold configuration file.

5. The method according to claim 1, characterized in that, The step of preprocessing the original partial discharge signal to obtain the target partial discharge signal includes: According to the preset cutoff frequency, the original partial discharge signal is subjected to anti-aliasing filtering to obtain the filtered partial discharge signal. The filtered partial discharge signal is amplified according to the target gain, and the amplified partial discharge signal is then converted from analog to digital to obtain the target partial discharge signal.

6. The method according to any one of claims 1 to 5, characterized in that, The method further includes: Based on the false alarm rate and the false alarm rate of the partial discharge of the generator stator, a feedback signal is generated, and based on the feedback signal, the stochastic gradient descent algorithm is driven to update the parameters, so as to obtain the updated stochastic gradient descent algorithm. The step of calculating the false alarm rate and false negative rate of partial discharge of the generator stator using a stochastic gradient descent algorithm based on the statistical characteristics, frequency domain characteristics, and discharge type includes: The statistical features, the frequency domain features, and the discharge type are input into the updated stochastic gradient descent algorithm to obtain the false alarm rate and the missed alarm rate of the generator stator partial discharge.

7. An optimization device for the detection threshold of partial discharge in generator stator, characterized in that, The device comprises: The signal processing module is used to acquire the original partial discharge signal collected by the partial discharge sensor of the generator under test, and to perform signal preprocessing on the original partial discharge signal to obtain the target partial discharge signal; The feature extraction module is used to extract features and recognize patterns from the target partial discharge signal to obtain the statistical features, frequency domain features, and discharge type of the generator stator partial discharge of the target partial discharge signal. The function determination module is used to calculate the false alarm rate and false negative rate of partial discharge of the generator stator using a stochastic gradient descent algorithm based on the statistical characteristics, the frequency domain characteristics and the discharge type, and to determine the comprehensive loss function based on the false alarm rate and the false negative rate. The threshold optimization module is used to obtain an initial detection threshold, and optimize the initial detection threshold according to the learning rate of the stochastic gradient descent algorithm and the comprehensive loss function to obtain an updated target detection threshold.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images