Partial discharge signal identification method and device and nonvolatile storage medium

Abstract

The invention discloses a partial discharge signal identification method and device and a nonvolatile storage medium. The method comprises the following steps: acquiring a real-time pulse signal of target electrical equipment; segmenting the real-time pulse signal, and extracting a plurality of feature vectors; performing quantum state coding on the plurality of feature vectors to obtain a plurality of initial quantum states; based on the multiple initial quantum states, a target quantum model is adopted for quantum evolution, and multiple target quantum states are obtained; respectively measuring the plurality of target quantum states to obtain a plurality of initial recognition results; repeating the steps of quantum coding, evolution and measurement for multiple times to obtain a plurality of initial recognition result groups; and respectively determining a target identification result in the plurality of initial identification result groups as a partial discharge identification result of the real-time pulse signal. The method solves the technical problems that a traditional partial discharge signal processing method is difficult to effectively extract signal features in a strong noise environment and cannot fully mine deep-level associated information of signals.

Description

Technical Field

[0001] This invention relates to the field of power system detection technology, and more specifically, to a method, apparatus, and non-volatile storage medium for identifying partial discharge signals. Background Technology

[0002] Partial discharge (PD) monitoring is a crucial aspect of power equipment condition assessment. Traditional techniques such as wavelet transform and empirical mode decomposition, while capable of noise reduction, suffer from weak adaptability and are prone to signal distortion or residue in noisy environments. Furthermore, feature selection relies on expert experience, limiting the full extraction of nonlinear and deep correlation information. As new power systems demand higher accuracy and real-time performance, the shortcomings of classical methods become increasingly apparent, particularly when processing high-dimensional, nonlinear PD signals, where intelligent feature extraction and model generalization capabilities become bottlenecks.

[0003] To address the limitations of traditional power PD signal processing techniques, existing deep learning methods, which construct deep neural networks to learn and classify features of PD signals, have significantly improved the automation level and classification accuracy of signal processing. Deep learning models perform particularly well on large-scale datasets, automatically extracting complex signal features. However, this approach still faces two major challenges: First, field data from power equipment is often limited, and annotation costs are high, restricting the small-sample adaptability and generalization ability of deep learning models; second, model training and inference are computationally intensive, especially in real-time monitoring scenarios, making it difficult to meet the demands of rapid response.

[0004] There is currently no effective solution to the above problems. Summary of the Invention

[0005] This invention provides a method, apparatus, and non-volatile storage medium for identifying partial discharge signals, thereby addressing the technical problem that traditional partial discharge signal processing methods struggle to effectively extract signal features in strong noise environments and fail to fully mine deep-level correlation information of signals.

[0006] According to one aspect of the present invention, a method for identifying partial discharge signals is provided, comprising: acquiring a real-time pulse signal of a target electrical device; segmenting the real-time pulse signal and extracting multiple feature vectors, wherein the feature vectors include multiple key features, including pulse amplitude, pulse width, rise time, zero-crossing rate, dominant frequency, centroid frequency, wavelet energy entropy, and sample entropy; performing quantum state encoding on the multiple feature vectors respectively to obtain multiple initial quantum states; and performing quantum evolution based on the multiple initial quantum states using a target quantum model respectively to obtain multiple target quantum states, wherein the target quantum model is trained based on the initial quantum model. The initial quantum model is a quantum perceptron model, which includes an input layer, a hidden layer, and an output layer. Multiple target quantum states are measured to obtain multiple initial identification results. Each initial identification result includes the type of a partial discharge signal included in the real-time pulse signal, such as corona discharge, internal discharge, free particle discharge, or suspended discharge. The quantum encoding, evolution, and measurement steps are repeated multiple times to obtain multiple sets of initial identification results for the real-time pulse signal. Each set of initial identification results includes multiple initial identification results. Finally, a target identification result is determined from each of the multiple initial identification result sets as the partial discharge identification result of the real-time pulse signal.

[0007] Optionally, the real-time pulse signal is segmented and multiple feature vectors are extracted, including: performing analog-to-digital conversion on the real-time pulse signal to obtain a digital signal of the real-time pulse signal; performing noise reduction on the digital signal to obtain a clean signal; performing pulse detection and segmentation on the clean signal to obtain a set of pulse waveforms, wherein the set of pulse waveforms contains multiple pulse segments; and performing feature extraction on the multiple pulse segments respectively to obtain multiple key features corresponding to each of the multiple pulse segments, which are used as multiple feature vectors.

[0008] Optionally, multiple eigenvectors are quantum state encoded to obtain multiple initial quantum states, including: normalizing multiple eigenvectors to obtain multiple standard eigenvectors; and encoding multiple initial quantum states based on the multiple standard eigenvectors.

[0009] Optionally, multiple target quantum states are measured to obtain multiple initial identification results, including: collapsing multiple target quantum states to obtain multiple initial qubit strings; obtaining the values ​​of preset qubit bits in multiple initial qubit strings to obtain multiple target binary strings; and determining multiple initial identification results based on multiple target binary strings.

[0010] Optionally, a target identification result is determined from multiple initial identification result groups as the partial discharge identification result of the real-time pulse signal, including: in each initial identification result group, the probability distribution of multiple initial identification results is statistically analyzed to obtain the target identification result, wherein the target identification result is the initial identification result with the highest probability among multiple initial identification results.

[0011] Optionally, the target quantum model is trained based on the initial quantum model, including: acquiring a partial discharge sample dataset, wherein the partial discharge sample dataset includes historical partial discharge data and corresponding type labels for the historical partial discharge data, the type labels including corona discharge, internal discharge, free particle discharge, and suspended discharge; constructing an initial quantum model; training the initial quantum model based on the partial discharge sample dataset and calculating the loss function value, wherein the loss function value is used to characterize the difference between the training result and the true value; repeating the quantum model training process, adjusting the target parameters based on the loss function value to obtain the target quantum model, wherein the target parameters are the operating parameters of the tunable quantum gate in the quantum circuit, and the target quantum model is a converged quantum perceptron model.

[0012] According to another aspect of the present invention, a partial discharge signal identification device is also provided, comprising: an acquisition module for acquiring a real-time pulse signal of a target electrical device; an extraction module for segmenting the real-time pulse signal and extracting multiple feature vectors, wherein the feature vectors include multiple key features, including pulse amplitude, pulse width, rise time, zero-crossing rate, dominant frequency, centroid frequency, wavelet energy entropy, and sample entropy; an encoding module for performing quantum state encoding on the multiple feature vectors respectively to obtain multiple initial quantum states; and an evolution module for performing quantum evolution based on the multiple initial quantum states using a target quantum model respectively to obtain multiple target quantum states, wherein the target quantum model is trained based on the initial quantum model. The initial quantum model is a quantum perceptron model, which includes an input layer, a hidden layer, and an output layer. A measurement module is used to measure multiple target quantum states to obtain multiple initial identification results. Each initial identification result includes the type of a partial discharge signal included in the real-time pulse signal, such as corona discharge, internal discharge, free particle discharge, or suspended discharge. A repetition module is used to repeat the above quantum encoding, evolution, and measurement steps multiple times to obtain multiple sets of initial identification results for the real-time pulse signal. Each set of initial identification results includes multiple initial identification results. A determination module is used to determine one target identification result from each of the multiple initial identification result sets as the partial discharge identification result of the real-time pulse signal.

[0013] According to another aspect of the present invention, a non-volatile storage medium is also provided, the non-volatile storage medium including a stored program, wherein, when the program is running, the device where the non-volatile storage medium is located is controlled to execute any of the above-described partial discharge signal identification methods.

[0014] According to another aspect of the present invention, a computer device is also provided, the computer device including a processor, the processor being configured to run a program, wherein the program executes any of the above-described partial discharge signal identification methods during runtime.

[0015] According to another aspect of the present invention, a computer program product is also provided, including a computer program that, when executed by a processor, implements any of the above-described partial discharge signal identification methods.

[0016] In this embodiment of the invention, a partial discharge signal identification method is employed. This involves acquiring the real-time pulse signal of the target electrical device; segmenting the real-time pulse signal and extracting multiple feature vectors, where each feature vector includes multiple key features such as pulse amplitude, pulse width, rise time, zero-crossing rate, dominant frequency, centroid frequency, wavelet energy entropy, and sample entropy; quantum state encoding is performed on each feature vector to obtain multiple initial quantum states; based on these initial quantum states, quantum evolution is performed using a target quantum model to obtain multiple target quantum states, where the target quantum model is trained based on the initial quantum model, which is a quantum perceptron model including an input layer, a hidden layer, and an output layer; and measurements are performed on each of the multiple target quantum states to obtain multiple initial identification results. Each initial identification result includes the type of a partial discharge signal included in the real-time pulse signal, such as corona discharge, internal discharge, free particle discharge, or suspended discharge. By repeating the quantum encoding, evolution, and measurement steps described above multiple times, multiple initial identification result groups of the real-time pulse signal are obtained. Each initial identification result group includes multiple initial identification results. A target identification result is determined from each of the multiple initial identification result groups as the partial discharge identification result of the real-time pulse signal. This achieves the goal of improving the stability and reliability of partial discharge signal identification, thereby realizing the technical effect of accurate and efficient condition monitoring and fault diagnosis of power equipment. Furthermore, it solves the technical problem that traditional partial discharge signal processing methods are difficult to effectively extract signal features in strong noise environments and cannot fully explore the deep-level correlation information of the signal. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0018] Figure 1 A hardware structure block diagram of a computer terminal for implementing a partial discharge signal identification method is shown.

[0019] Figure 2 This is a flowchart illustrating the partial discharge signal identification method provided according to an embodiment of the present invention;

[0020] Figure 3 This is a structural diagram of a quantum sensor model provided by an optional embodiment of the present invention;

[0021] Figure 4 This is an adaptive processor framework diagram provided according to an optional embodiment of the present invention;

[0022] Figure 5 This is a flowchart of partial discharge signal identification and processing provided by an optional embodiment of the present invention;

[0023] Figure 6 This is a structural block diagram of a partial discharge signal identification device provided according to an embodiment of the present invention. Detailed Implementation

[0024] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0025] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0026] According to an embodiment of the present invention, a method for identifying partial discharge signals is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0027] The method embodiment provided in Embodiment 1 of this application can be executed on a mobile terminal, computer terminal, or similar computing device. Figure 1 A hardware block diagram of a computer terminal for implementing a partial discharge signal identification method is shown. Figure 1 As shown, the computer terminal 10 may include one or more processors (shown as 102a, 102b, ..., 102n in the figure) (the processor may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, computer terminal 10 may also include... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0028] It should be noted that the aforementioned one or more processors and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 10. As involved in the embodiments of this application, the data processing circuits serve as a processor control mechanism (e.g., selection of a variable resistor termination path connected to an interface).

[0029] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the partial discharge signal identification method in this embodiment of the invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby realizing the partial discharge signal identification method of the aforementioned application. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor, and these remote memories can be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0030] The display can be, for example, a touchscreen liquid crystal display (LCD) that allows the user to interact with the user interface of the computer terminal 10.

[0031] Figure 2 This is a flowchart illustrating the partial discharge signal identification method provided in an embodiment of the present invention, as shown below. Figure 2 As shown, the method includes the following steps:

[0032] Step S201: Obtain the real-time pulse signal of the target electrical equipment.

[0033] In this step, sensors deployed at critical locations in the power system (such as high-frequency current sensors (HFCT), ultra-high frequency sensors (UHF), ultrasonic sensors, or transient ground voltage sensors) capture partial discharge (PD) signals related to the insulation status of the equipment in real time. The sensors convert changes in electromagnetic or acoustic waves in the environment into recordable pulse signals, which reflect potential partial discharge activity within the target electrical equipment.

[0034] Signal acquisition should consider the real-time and continuous operation of power equipment; that is, sensors should continuously monitor the status of power equipment to detect abnormal discharge activities in a timely manner. Furthermore, considering the complexity and diversity of actual operation, the acquired real-time pulse signals may contain various signal forms from different discharge types, as well as various noises from equipment operation and the environment. Subsequent noise reduction processing is required to accurately identify and classify these signals.

[0035] Step S202: The real-time pulse signal is segmented and multiple feature vectors are extracted. The feature vectors include multiple key features, including pulse amplitude, pulse width, rise time, zero crossing rate, main frequency, centroid frequency, wavelet energy entropy, and sample entropy.

[0036] In this step, the real-time pulse signal is first converted from analog to digital (ADC) format to facilitate further computer analysis. Then, adaptive noise reduction is applied to the digital signal to effectively reduce accompanying electromagnetic interference and background noise, ensuring the accuracy of subsequent feature extraction. Further, the denoised signal is segmented using peak detection and fixed window techniques to cut the signal sequence into segments containing complete partial discharge pulses. By analyzing each pulse segment, a series of key features are extracted to form a feature vector, which describes the specific properties of the partial discharge.

[0037] Specifically, key features can include pulse amplitude, pulse width, rise time, zero-crossing rate, dominant frequency, centroid frequency, wavelet energy entropy, and sample entropy. Pulse amplitude is the maximum amplitude of the signal, reflecting the magnitude of partial discharge energy; pulse width refers to the length of time the signal remains above a specific threshold, helping to distinguish the duration of different types of discharges; rise time is the time required for the signal amplitude to rise from 10% to 90%, which can be used to analyze the transient characteristics of the signal; zero-crossing rate can be used to measure the frequency at which the signal crosses zero; dominant frequency is the frequency with the highest energy in the signal spectrum calculated by Fast Fourier Transform (FFT); centroid frequency, also based on spectral analysis, calculates the average frequency of the signal spectrum, providing information about the central trend of the signal's frequency component distribution; wavelet energy entropy refers to the signal energy distribution at different scales obtained through wavelet transform, and then calculating its entropy value, reflecting the complexity and information content of the signal in the time-frequency domain, particularly suitable for the analysis of non-stationary signals; sample entropy is obtained by calculating the number of similar patterns and their uncertainty in the signal, and can be used to quantify the randomness and complexity of signal sequences. These features together constitute a comprehensive description of the pulse signal, capable of capturing the time-domain, frequency-domain, and nonlinear characteristics of the partial discharge signal, providing crucial information for subsequent quantum state encoding and classification.

[0038] Step S203: Encode multiple eigenvectors into quantum states to obtain multiple initial quantum states.

[0039] In this step, multiple eigenvectors extracted from the preprocessed signal are mapped to quantum states. Specifically, for each eigenvector, its elements contain various key properties of the partial discharge pulse, such as pulse amplitude and pulse width. To convert these classical values ​​into a form operable by quantum computing, each eigenvalue needs to be normalized to ensure it falls within the applicable range of the quantum gate control parameters. Then, an angle encoding strategy is employed, mapping each normalized eigenvalue to the angle of the Ry rotation gate on the qubit. The initial quantum state |ψ> is then calculated.

[0040] Step S204: Based on multiple initial quantum states, quantum evolution is performed using target quantum models to obtain multiple target quantum states. The target quantum models are trained based on the initial quantum models, which are quantum perceptron models. The quantum perceptron model includes an input layer, a hidden layer, and an output layer.

[0041] In this step, a quantum perceptron model is constructed, and the encoded initial quantum state is used as the input quantum state |ψ_in> and subjected to a series of controlled, non-classical unitary evolutions to transform it into an output quantum state |ψ_out>, which is the target quantum state. Figure 3 This is a structural diagram of a quantum perceptron model provided by an optional embodiment of the present invention, such as... Figure 3 As shown, the quantum perceptron model includes an input layer, one or more hidden layers consisting of parameterized quantum gates, and an output layer.

[0042] Specifically, the encoded initial quantum state |ψ> is input into this parameterized quantum circuit. The circuit consists of a series of preset fixed quantum gates and trainable parameterized quantum gates. By applying these quantum gates, the evolution of the input quantum state |ψ> is achieved, and the evolution formula is as follows:

[0043]

[0044] Where U(θ) is the unitary transformation of the entire parameterized quantum circuit, and θ is the set of trainable parameters.

[0045] The output is the final quantum state |ψ_out> after evolution by the quantum perceptron. This quantum state is the result of complex, nonlinear transformations of the input features in quantum space, and its probability amplitude distribution already contains the model's judgment on the type of partial discharge. This quantum state will be used as the target quantum state for further measurement to read out the identification result of the partial discharge signal.

[0046] Step S205: Measure multiple target quantum states to obtain multiple initial identification results. The initial identification results include the type of a partial discharge signal included in the real-time pulse signal, which may be corona discharge, internal discharge, free particle discharge, or suspension discharge.

[0047] In this step, quantum measurement operations are used to transform the information in the quantum state into a interpretable discharge type discrimination result. Each target quantum state, after undergoing a quantum evolution process, carries quantum information related to the partial discharge type. Quantum measurement is performed at the output stage of quantum computing; for the output quantum state of the target quantum model, the state of the qubit is measured under the computational basis.

[0048] Specifically, measuring the state of a qubit requires focusing on the output qubits defined in the model structure. The number and position of these qubits depend on the number of discharge types to be classified. In this embodiment, since partial discharge types include four types—corona discharge, internal discharge, free particle discharge, and levitation discharge—the first two qubits are selected for measurement. The measurement result is a classical binary bit string, such as "00", "01", "10", or "11", corresponding to the four discharge types mentioned above.

[0049] Step S206: Repeat the above steps of quantum encoding, evolution and measurement multiple times to obtain multiple initial identification result groups of real-time pulse signals, wherein the initial identification result group includes multiple initial identification results.

[0050] In this step, in order to extract stable information from the quantum state, multiple rounds of quantum processing are required on the same or similar input real-time pulse signal. By repeatedly executing the quantum encoding, evolution and measurement process, the accuracy and robustness of discharge type identification can be improved.

[0051] Specifically, with the operating parameter θ fixed, the aforementioned quantum encoding, evolution, and measurement process needs to be repeated multiple times. Each time, the same or similar real-time pulse signal is used as input, and an initial identification result is obtained through an independent quantum computing process. Based on this, for each real-time pulse signal, an initial identification result set consisting of multiple measurement results is obtained. By statistically analyzing the frequency of occurrence of different discharge type identification results in each result set, a probability distribution can be derived. Then, based on this distribution, the most likely discharge type is determined as the final identification result. This method overcomes the randomness and uncertainty of a single measurement and ensures the robustness of the identification results.

[0052] Step S207: Determine one target identification result from each of the multiple initial identification result groups as the partial discharge identification result of the real-time pulse signal.

[0053] In this step, the initial identification results in each result group are statistically processed to calculate the frequency of occurrence of each discharge type, thereby obtaining its empirical probability distribution. Based on this probability distribution, the discharge type with the highest probability is selected as the target identification result, forming the final judgment on the partial discharge type in the real-time pulse signal. The above-described probabilistic statistical decision-making method fully considers the randomness of quantum measurement and the importance of multiple experiments, reducing the uncertainty of a single measurement and ensuring that even in a strong noise environment, a more robust and accurate partial discharge type identification can be made based on the results of multiple experiments.

[0054] As an optional embodiment, the real-time pulse signal is segmented and multiple feature vectors are extracted, including: performing analog-to-digital conversion on the real-time pulse signal to obtain a digital signal of the real-time pulse signal; performing noise reduction processing on the digital signal to obtain a clean signal; performing pulse detection and segmentation on the clean signal to obtain a set of pulse waveforms, wherein the set of pulse waveforms contains multiple pulse segments; and performing feature extraction on the multiple pulse segments respectively to obtain multiple key features corresponding to each of the multiple pulse segments, which serve as multiple feature vectors.

[0055] Optionally, the acquired real-time pulse signal is x(t), which is an analog signal. After analog-to-digital conversion (ADC), a discrete digital signal sequence {x_n} can be obtained. In addition to the partial discharge signal, the real-time pulse signal may sometimes include strong periodic sinusoidal interference. An adaptive processor is used to perform preliminary filtering on the real-time pulse signal. Based on the characteristics of narrowband interference during signal detection, the filter characteristics are automatically adjusted to suppress some background interference and retain the main shape of the PD pulse.

[0056] Specifically, Figure 4 This is an adaptive processor framework diagram provided according to an optional embodiment of the present invention. Figure 4 As shown, the adaptive processor includes a digital filter h, an adder ∑, and an adaptive algorithm. The main input is y(t), where y(t) = p(t) + n(t), p(t) is the useful real partial discharge pulse signal, and n(t) is the interference signal superimposed on the partial discharge signal. The reference input n'(t) is a noise signal that is highly correlated with n(t) but uncorrelated with p(t). (t) is the output of h, when the square of the adder output is e 2 When (t) is not minimum, the filter is modified according to the given adaptive algorithm to make e 2 (t) decreases. Adjust the filter h. (t) is made close to n(t), and finally the output e(t) of the adder is close to p(t), thus obtaining a pure signal {p_n}.

[0057] Subsequently, pulse detection is performed on the denoised signal {p_n}. Pulse segmentation is performed using fixed threshold detection based on amplitude thresholds or waveform characteristics. A segment of the denoised signal is then observed, and analysis is conducted based on experience or statistical analysis of signal peak values.

[0058]

[0059] Wherein, β is an empirical coefficient between 0.2 and 0.4.

[0060] A pulse is considered detected when the absolute value of the amplitude p_i at any data point, |p_i|, exceeds a fixed threshold V_th. Once a pulse is detected, fixed lengths M points are directly truncated forward and backward from that point (peak point). All truncated pulse segments form a set p_j[n] of partial discharge pulse waveforms that are naturally aligned with the peak points.

[0061] Finally, a set of multidimensional features is extracted from each segmented discharge pulse to form a multidimensional feature vector V_c, where V_c=[f1,f2,f3,...,f_d] is a d-dimensional feature vector. In this optional embodiment, eight features, such as pulse amplitude, are selected as key features. The feature vector can be specifically represented as V_c=[pulse amplitude, pulse width, rise time, zero-crossing rate, dominant frequency, centroid frequency, wavelet energy entropy, sample entropy].

[0062] As an optional embodiment, multiple eigenvectors are quantum state encoded to obtain multiple initial quantum states, including: normalizing multiple eigenvectors to obtain multiple standard eigenvectors; and encoding multiple initial quantum states based on the multiple standard eigenvectors.

[0063] Optionally, the d-dimensional feature vector V_c=[f1,f2,..,f_d] is normalized, and each feature component is normalized to the interval [0,1] to obtain the standard feature vector V_c'=[f ,f ,...,f'_d], to ensure that it satisfies the normalization constraint of the quantum state or the domain of the rotation angle.

[0064] Secondly, using angular encoding, each component f'_i of the normalized standard feature vector V_c' is transformed by a linear transformation θ_i=π. f'i is mapped to the Ry rotation gate angle of the i-th qubit. The initial quantum state |ψ> is prepared by applying Ry(θ_i) gates in parallel on all d qubits, as shown in the following formula:

[0065]

[0066] Here, the output |ψ> is a prepared initial quantum state characterizing the partial discharge features of the input. This quantum state will be used as the input to the target quantum model and fed into the subsequent quantum perceptron for evolution and computation.

[0067] As an optional embodiment, multiple target quantum states are measured to obtain multiple initial identification results, including: collapsing multiple target quantum states to obtain multiple initial qubit strings; obtaining the values ​​of preset qubit bits in multiple initial qubit strings to obtain multiple target binary strings; and determining multiple initial identification results based on multiple target binary strings.

[0068] Optionally, the classification information of the target quantum state is first collapsed into a defined classical bit string, and the final partial discharge type discrimination result is generated based on this. The measurement result of each quantum bit (|0> or |1>) constitutes a binary string, based on the defined mapping relationship f:{0,1}^m→{ , The formula ,...,C_K} can map the m-bit binary output to a specific discharge type, and use the measured binary string as the partial discharge identification result.

[0069] Specifically, in this optional embodiment, the first two qubits are selected for measurement, resulting in four binary strings, each corresponding to one of four identification types. The binary string output "00" represents corona discharge, "01" represents internal discharge, "10" represents free particle discharge, and "11" represents levitation discharge. By performing the above measurements on the target quantum state multiple times, multiple initial identification results can be obtained. These multiple measurement results for the target quantum state constitute multiple sets of initial identification results.

[0070] As an optional embodiment, a target identification result is determined from multiple initial identification result groups as the partial discharge identification result of the real-time pulse signal, including: in each initial identification result group, the probability distribution of multiple initial identification results is statistically analyzed to obtain the target identification result, wherein the target identification result is the initial identification result with the highest probability of occurrence among multiple initial identification results.

[0071] Optionally, since quantum measurement is inherently probabilistic, a single measurement result is random and cannot reliably represent the model's final judgment. Therefore, a statistical decision-making mechanism must be employed to determine the target identification result. Specifically, in each initial identification result group, the frequency of occurrence of each possible identification type is counted, and the output frequency of each discharge type (corona discharge, internal discharge, free particle discharge, and suspension discharge) is captured, thereby constructing a probability distribution based on the initial identification results of that group. Secondly, based on the obtained probability distribution, the empirical probability of the identification result is calculated, and the output type with the highest frequency and largest empirical probability is identified. This is used as the final identification result for the partial discharge phenomenon in the real-time pulse signal, ensuring the robustness and accuracy of the identification process.

[0072] As an optional implementation, the target quantum model is trained based on the initial quantum model, including: acquiring a partial discharge sample dataset, wherein the partial discharge sample dataset includes historical partial discharge data and corresponding type labels for the historical partial discharge data, the type labels including corona discharge, internal discharge, free particle discharge, and suspended discharge; constructing an initial quantum model; training the initial quantum model based on the partial discharge sample dataset and calculating the loss function value, wherein the loss function value is used to characterize the difference between the training result and the true value; repeating the quantum model training process, adjusting the target parameters based on the loss function value to obtain the target quantum model, wherein the target parameters are the operating parameters of the tunable quantum gate in the quantum circuit, and the target quantum model is a converged quantum perceptron model.

[0073] Optionally, firstly, a partial discharge sample dataset is obtained, consisting of historical partial discharge data and its type labels, containing 4000 samples. The historical partial discharge data covers previously recorded instances of corona discharge, internal discharge, free particle discharge, and suspended discharge, while the type labels clearly identify the specific discharge type to which each data point belongs.

[0074] Subsequently, an initial quantum model is constructed, namely an untrained quantum perceptron model, which consists of an input layer, a hidden layer composed of a series of parameterized quantum gates, and an output layer. The input layer is responsible for receiving the encoded quantum state, while the parameterized quantum gates allow the model to adapt to the characteristic patterns of different types of partial discharges by adjusting the parameters.

[0075] Furthermore, based on a partial discharge sample dataset, the quantum perceptron model is iteratively trained. A loss function is introduced to quantify the difference between the model's predictions and the true type labels of the samples. The model training process is repeated, and in each iteration, the operating parameters of the tunable quantum gates in the quantum circuit, i.e., the target parameters, are adjusted based on the calculated loss function value. This process utilizes classical optimization algorithms to gradually approach the optimal parameter settings, making the model's predictions closer to the actual discharge types. Iterative training continues until the model converges. The resulting quantum perceptron model is the target quantum model, which can accurately and efficiently identify and classify partial discharges, demonstrating significant diagnostic performance.

[0076] As an optional embodiment, Figure 5 This is a flowchart of partial discharge signal identification and processing provided according to an optional embodiment of the present invention. Figure 5As shown, signal acquisition and preprocessing are first performed. High-frequency current sensors (HFCT), ultra-high frequency sensors (UHF), ultrasonic sensors, or transient ground voltage sensors capture partial discharge signals at the monitoring points of electrical equipment. Subsequently, the analog signals are converted into digital signal sequences for subsequent processing via analog-to-digital converters (ADCs). The digital signal sequences are filtered by an adaptive processor to remove periodic sinusoidal interference and background noise, retaining only pure partial discharge pulse signals. Then, pulse segmentation and alignment are performed on each pulse signal. A V_th threshold is used to detect pulses, normalizing and aligning signal segments for unified analysis. Further, the time-domain, frequency-domain, time-frequency-domain, and nonlinear features extracted from each pulse are integrated into a classical feature vector, serving as the basis for quantum state encoding. Quantum state encoding maps the classical feature vector to quantum states, using angle encoding and Ry rotation gates to encode the feature values ​​as the rotation angle of the qubits, achieving quantum representation of the features. Finally, the encoded quantum state is fed into a quantum circuit, undergoing a series of controlled unitary evolutions. Through the adjustment of parameterized quantum gates, the quantum state evolves to a state that reflects the type of partial discharge. This stage includes the input encoding layer, the hidden layer composed of parameterized quantum gates, and the final evolution of the quantum state. Finally, the evolved quantum state is measured, the quantum bit states corresponding to each discharge type are statistically analyzed, and based on the statistical analysis of the measurement results, the partial discharge type is determined, thus classifying the signal.

[0077] Through the above steps, the stability and reliability of partial discharge signal identification are improved, thereby achieving the technical effect of accurate and efficient condition monitoring and fault diagnosis of power equipment. This solves the technical problem that traditional partial discharge signal processing methods are unable to effectively extract signal features in strong noise environments and cannot fully explore the deep-level correlation information of signals.

[0078] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0079] Through the above description of the embodiments, those skilled in the art can clearly understand that the partial discharge signal identification method according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platform. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0080] According to embodiments of the present invention, an apparatus for implementing the above-described partial discharge signal identification method is also provided. Figure 6 This is a structural block diagram of a partial discharge signal identification device provided according to an embodiment of the present invention, such as... Figure 6 As shown, the device includes: an acquisition module 61, an extraction module 62, an encoding module 63, an evolution module 64, a measurement module 65, a repetition module 66, and a determination module 67. The device will be described below.

[0081] Acquisition module 61 is used to acquire the real-time pulse signal of the target electrical equipment.

[0082] Extraction module 62, connected to acquisition module 61, is used to segment the real-time pulse signal and extract multiple feature vectors. The feature vectors include multiple key features, including pulse amplitude, pulse width, rise time, zero crossing rate, main frequency, centroid frequency, wavelet energy entropy, and sample entropy.

[0083] The encoding module 63, connected to the extraction module 62, is used to encode multiple feature vectors into quantum states to obtain multiple initial quantum states.

[0084] Evolution module 64, connected to encoding module 63, is used to perform quantum evolution based on multiple initial quantum states using target quantum models to obtain multiple target quantum states. The target quantum model is trained based on the initial quantum model, which is a quantum perceptron model. The quantum perceptron model includes an input layer, a hidden layer, and an output layer.

[0085] The measurement module 65, connected to the evolution module 64, is used to measure multiple target quantum states respectively and obtain multiple initial identification results. The initial identification results include the type of a partial discharge signal included in the real-time pulse signal, which may be corona discharge, internal discharge, free particle discharge or suspension discharge.

[0086] The repeat module 66, connected to the measurement module 65, is used to repeat the above-mentioned quantum encoding, evolution and measurement steps multiple times to obtain multiple initial identification result groups of real-time pulse signals, wherein the initial identification result group includes multiple initial identification results.

[0087] The determination module 67, connected to the repeating module 66, is used to determine a target identification result from multiple initial identification result groups as the partial discharge identification result of the real-time pulse signal.

[0088] It should be noted that the acquisition module 61, extraction module 62, encoding module 63, evolution module 64, measurement module 65, repetition module 66, and determination module 67 mentioned above correspond to steps S201 to S207 in the embodiments. Multiple modules and their corresponding steps implement the same instances and application scenarios, but are not limited to the content disclosed in the above embodiments. It should also be noted that the above modules, as part of the device, can run on the computer terminal 10 provided in the embodiments.

[0089] Embodiments of the present invention may provide a computer device. Optionally, in this embodiment, the computer device may be located in at least one of a plurality of network devices in a computer network. The computer device includes a memory and a processor.

[0090] The memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the partial discharge signal identification method and device in this embodiment of the invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, thereby realizing the aforementioned partial discharge signal identification method. The memory may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include memory remotely located relative to the processor, and these remote memories can be connected to a computer terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0091] The processor can access information and application programs stored in memory via a transmission device to perform the following steps: acquiring the real-time pulse signal of the target electrical device; segmenting the real-time pulse signal and extracting multiple feature vectors, where each feature vector includes multiple key features, such as pulse amplitude, pulse width, rise time, zero-crossing rate, dominant frequency, centroid frequency, wavelet energy entropy, and sample entropy; encoding each feature vector into quantum states to obtain multiple initial quantum states; and performing quantum evolution based on the multiple initial quantum states using the target quantum model to obtain multiple target quantum states, where the target quantum model is trained based on the initial quantum model. The initial quantum model is a quantum perceptron model, which includes an input layer, a hidden layer, and an output layer. Multiple target quantum states are measured to obtain multiple initial identification results. Each initial identification result includes the type of a partial discharge signal included in the real-time pulse signal, such as corona discharge, internal discharge, free particle discharge, or suspended discharge. The quantum encoding, evolution, and measurement steps are repeated multiple times to obtain multiple sets of initial identification results for the real-time pulse signal. Each set of initial identification results includes multiple initial identification results. Finally, a target identification result is determined from each of the multiple initial identification result sets as the partial discharge identification result of the real-time pulse signal.

[0092] Optionally, the processor may also execute program code for the following steps: segmenting the real-time pulse signal and extracting multiple feature vectors, including: performing analog-to-digital conversion on the real-time pulse signal to obtain a digital signal of the real-time pulse signal; performing noise reduction on the digital signal to obtain a clean signal; performing pulse detection and segmentation on the clean signal to obtain a set of pulse waveforms, wherein the set of pulse waveforms contains multiple pulse segments; and performing feature extraction on the multiple pulse segments respectively to obtain multiple key features corresponding to each of the multiple pulse segments, which serve as multiple feature vectors.

[0093] Optionally, the processor may also execute program code that performs the following steps: quantum state encoding on multiple eigenvectors to obtain multiple initial quantum states, including: normalizing multiple eigenvectors to obtain multiple standard eigenvectors; and performing quantum state encoding on multiple standard eigenvectors to obtain multiple initial quantum states.

[0094] Optionally, the processor may also execute program code that performs the following steps: measuring multiple target quantum states to obtain multiple initial identification results, including: collapsing multiple target quantum states to obtain multiple initial qubit strings; obtaining the values ​​of preset qubit bits in multiple initial qubit strings to obtain multiple target binary strings; and determining multiple initial identification results based on multiple target binary strings.

[0095] Optionally, the processor may also execute program code that performs the following steps: determining a target identification result from multiple initial identification result groups as the partial discharge identification result of the real-time pulse signal, including: statistically analyzing the probability distribution of multiple initial identification results in each initial identification result group to obtain the target identification result, wherein the target identification result is the initial identification result with the highest probability of occurrence among the multiple initial identification results.

[0096] Optionally, the processor may also execute program code for the following steps: the target quantum model is trained based on the initial quantum model, including: acquiring a partial discharge sample dataset, wherein the partial discharge sample dataset includes historical partial discharge data and type labels corresponding to the historical partial discharge data, the type labels including corona discharge, internal discharge, free particle discharge, and suspended discharge; constructing an initial quantum model; training the initial quantum model based on the partial discharge sample dataset and calculating the loss function value, wherein the loss function value is used to characterize the difference between the training result and the true value; repeating the quantum model training process, adjusting the target parameters based on the loss function value to obtain the target quantum model, wherein the target parameters are the operating parameters of the tunable quantum gate in the quantum circuit, and the target quantum model is a converged quantum perceptron model.

[0097] This invention provides a scheme for partial discharge signal identification. The scheme involves acquiring real-time pulse signals from a target electrical device; segmenting the real-time pulse signals and extracting multiple feature vectors, where each feature vector includes multiple key features such as pulse amplitude, pulse width, rise time, zero-crossing rate, dominant frequency, centroid frequency, wavelet energy entropy, and sample entropy; encoding each feature vector into quantum states to obtain multiple initial quantum states; and performing quantum evolution using a target quantum model based on these initial quantum states to obtain multiple target quantum states. The target quantum model is trained based on the initial quantum model, which is a quantum perceptron model comprising an input layer, a hidden layer, and an output layer. Finally, the multiple target quantum states are measured to obtain multiple initial identification signals. The process involves several steps. The initial identification result includes the type of partial discharge signal included in the real-time pulse signal, such as corona discharge, internal discharge, free particle discharge, or suspended discharge. The quantum encoding, evolution, and measurement steps are repeated multiple times to obtain multiple initial identification result groups for the real-time pulse signal. Each initial identification result group contains multiple initial identification results. A target identification result is then determined from each of these initial identification result groups as the partial discharge identification result for the real-time pulse signal. This improves the stability and reliability of partial discharge signal identification, thereby solving the technical problem that traditional partial discharge signal processing methods struggle to effectively extract signal features in noisy environments and cannot fully mine deep-level correlation information in the signal.

[0098] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a non-volatile storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0099] Embodiments of the present invention also provide a non-volatile storage medium. Optionally, in this embodiment, the aforementioned non-volatile storage medium can be used to store the program code executed by the partial discharge signal identification method provided in the above embodiments.

[0100] Optionally, in this embodiment, the non-volatile storage medium may be located in any computer terminal in a group of computer terminals in a computer network, or in any mobile terminal in a group of mobile terminals.

[0101] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: acquiring the real-time pulse signal of the target electrical device; segmenting the real-time pulse signal and extracting multiple feature vectors, wherein the feature vectors include multiple key features, including pulse amplitude, pulse width, rise time, zero-crossing rate, dominant frequency, centroid frequency, wavelet energy entropy, and sample entropy; quantum-state encoding the multiple feature vectors respectively to obtain multiple initial quantum states; and performing quantum evolution based on the multiple initial quantum states using the target quantum model respectively to obtain multiple target quantum states, wherein the target quantum model is trained based on the initial quantum model. The initial quantum model is a quantum perceptron model, which includes an input layer, a hidden layer, and an output layer. Multiple target quantum states are measured to obtain multiple initial identification results. Each initial identification result includes the type of a partial discharge signal included in the real-time pulse signal, such as corona discharge, internal discharge, free particle discharge, or suspended discharge. The quantum encoding, evolution, and measurement steps are repeated multiple times to obtain multiple sets of initial identification results for the real-time pulse signal. Each set of initial identification results includes multiple initial identification results. Finally, a target identification result is determined from each of the multiple initial identification result sets as the partial discharge identification result of the real-time pulse signal.

[0102] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: segmenting the real-time pulse signal and extracting multiple feature vectors, including: performing analog-to-digital conversion on the real-time pulse signal to obtain a digital signal of the real-time pulse signal; performing noise reduction processing on the digital signal to obtain a clean signal; performing pulse detection and segmentation on the clean signal to obtain a pulse waveform set, wherein the pulse waveform set contains multiple pulse segments; and performing feature extraction on the multiple pulse segments respectively to obtain multiple key features corresponding to each of the multiple pulse segments, which serve as multiple feature vectors.

[0103] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: quantum state encoding of multiple feature vectors to obtain multiple initial quantum states, including: normalizing multiple feature vectors to obtain multiple standard feature vectors; and quantum state encoding based on multiple standard feature vectors to obtain multiple initial quantum states.

[0104] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: measuring multiple target quantum states to obtain multiple initial identification results, including: collapsing multiple target quantum states to obtain multiple initial qubit strings; obtaining the values ​​of preset qubit bits in the multiple initial qubit strings to obtain multiple target binary strings; and determining multiple initial identification results based on the multiple target binary strings.

[0105] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: determining a target identification result in each of the multiple initial identification result groups as the partial discharge identification result of the real-time pulse signal, including: statistically analyzing the probability distribution of the multiple initial identification results in each initial identification result group to obtain the target identification result, wherein the target identification result is the initial identification result with the highest probability of occurrence among the multiple initial identification results.

[0106] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: the target quantum model is trained based on the initial quantum model, including: acquiring a partial discharge sample dataset, wherein the partial discharge sample dataset includes historical partial discharge data and type labels corresponding to the historical partial discharge data, the type labels including corona discharge, internal discharge, free particle discharge, and suspended discharge; constructing an initial quantum model; training the initial quantum model based on the partial discharge sample dataset and calculating the loss function value, wherein the loss function value is used to characterize the difference between the training result and the true value; repeating the quantum model training process, adjusting the target parameters based on the loss function value to obtain the target quantum model, wherein the target parameters are the operating parameters of the tunable quantum gate in the quantum circuit, and the target quantum model is a converged quantum perceptron model.

[0107] Embodiments of the present invention also provide a computer program product, including a computer program. Optionally, in this embodiment, when the computer program is executed by a processor, it can: acquire real-time pulse signals of a target electrical device; segment the real-time pulse signals and extract multiple feature vectors, wherein the feature vectors include multiple key features, including pulse amplitude, pulse width, rise time, zero-crossing rate, dominant frequency, centroid frequency, wavelet energy entropy, and sample entropy; encode the multiple feature vectors into quantum states to obtain multiple initial quantum states; and perform quantum evolution based on the multiple initial quantum states using a target quantum model to obtain multiple target quantum states, wherein the target quantum model is based on... The initial quantum model is trained and is a quantum perceptron model, which includes an input layer, a hidden layer, and an output layer. Multiple target quantum states are measured to obtain multiple initial identification results. Each initial identification result includes the type of a partial discharge signal included in the real-time pulse signal, such as corona discharge, internal discharge, free particle discharge, or suspended discharge. The quantum encoding, evolution, and measurement steps are repeated multiple times to obtain multiple sets of initial identification results for the real-time pulse signal. Each set of initial identification results includes multiple initial identification results. Finally, a target identification result is determined from each of the multiple initial identification result sets as the partial discharge identification result of the real-time pulse signal.

[0108] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0109] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0110] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0111] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0112] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0113] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a non-volatile storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0114] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for identifying partial discharge signals, characterized in that, include: Acquire the real-time pulse signal of the target electrical equipment; The real-time pulse signal is segmented and multiple feature vectors are extracted. The feature vectors include multiple key features, including pulse amplitude, pulse width, rise time, zero crossing rate, main frequency, centroid frequency, wavelet energy entropy, and sample entropy. Quantum state encoding is performed on the multiple feature vectors to obtain multiple initial quantum states; Based on the multiple initial quantum states, quantum evolution is performed using target quantum models to obtain multiple target quantum states. The target quantum models are trained based on the initial quantum models, which are quantum perceptron models, including an input layer, a hidden layer, and an output layer. The multiple target quantum states are measured respectively to obtain multiple initial identification results. The initial identification results include the type of a partial discharge signal included in the real-time pulse signal. The type includes corona discharge, internal discharge, free particle discharge or suspension discharge. By repeating the above steps of quantum encoding, evolution, and measurement multiple times, multiple initial identification result groups of the real-time pulse signal are obtained, wherein each initial identification result group includes multiple initial identification results; One target identification result is determined from each of the multiple initial identification result groups and used as the partial discharge identification result of the real-time pulse signal.

2. The method according to claim 1, characterized in that, The process of segmenting the real-time pulse signal and extracting multiple feature vectors includes: The real-time pulse signal is converted from analog to digital to obtain the digital signal of the real-time pulse signal; The digital signal is subjected to noise reduction processing to obtain a clean signal; The pure signal is subjected to pulse detection and segmentation to obtain a set of pulse waveforms, wherein the set of pulse waveforms contains multiple pulse segments; Feature extraction is performed on the multiple pulse segments respectively to obtain the multiple key features corresponding to each of the multiple pulse segments, which are used as the multiple feature vectors.

3. The method according to claim 1, characterized in that, The process of quantum state encoding of the multiple feature vectors to obtain multiple initial quantum states includes: The multiple feature vectors are normalized to obtain multiple standard feature vectors; Based on the multiple standard feature vectors, quantum state encoding is performed respectively to obtain the multiple initial quantum states.

4. The method according to claim 1, characterized in that, The measurement of the multiple target quantum states yields multiple initial identification results, including: The multiple target quantum states are collapsed to obtain multiple initial qubit strings; The values ​​of preset qubits in the multiple initial qubit strings are obtained respectively to obtain multiple target binary strings; Based on the multiple target binary strings, the multiple initial recognition results are determined respectively.

5. The method according to claim 1, characterized in that, The step of determining a target identification result from the plurality of initial identification result groups as the partial discharge identification result of the real-time pulse signal includes: In each initial identification result group, the probability distribution of multiple initial identification results is statistically analyzed to obtain the target identification result, wherein the target identification result is the initial identification result with the highest probability among the multiple initial identification results.

6. The method according to claim 1, characterized in that, The target quantum model is obtained by training an initial quantum model and includes: Obtain a partial discharge sample dataset, wherein the partial discharge sample dataset includes historical partial discharge data and type labels corresponding to the historical partial discharge data, and the type labels include corona discharge, internal discharge, free particle discharge and suspended discharge; Construct the initial quantum model; Based on the partial discharge sample dataset, the initial quantum model is trained, and the loss function value is calculated, wherein the loss function value is used to characterize the difference between the training result and the true value; The quantum model training process is repeated, and the target parameters are adjusted based on the loss function value to obtain the target quantum model. The target parameters are the operating parameters of the tunable quantum gate in the quantum circuit, and the target quantum model is the converged quantum perceptron model.

7. A partial discharge signal identification device, characterized in that, include: The acquisition module is used to acquire the real-time pulse signal of the target electrical equipment; The extraction module is used to segment the real-time pulse signal and extract multiple feature vectors, wherein the feature vectors include multiple key features, including pulse amplitude, pulse width, rise time, zero crossing rate, main frequency, centroid frequency, wavelet energy entropy, and sample entropy. The encoding module is used to encode the multiple feature vectors into quantum states to obtain multiple initial quantum states; An evolution module is used to perform quantum evolution based on the multiple initial quantum states using target quantum models to obtain multiple target quantum states. The target quantum models are trained based on the initial quantum models, which are quantum perceptron models, including an input layer, a hidden layer, and an output layer. The measurement module is used to measure the multiple target quantum states respectively and obtain multiple initial identification results. The initial identification results include the type of a partial discharge signal included in the real-time pulse signal. The type includes corona discharge, internal discharge, free particle discharge or suspension discharge. The repetition module is used to repeat the above-described quantum encoding, evolution, and measurement steps multiple times to obtain multiple initial identification result groups of the real-time pulse signal, wherein the initial identification result group includes multiple initial identification results; The determination module is used to determine a target identification result from the plurality of initial identification result groups as the partial discharge identification result of the real-time pulse signal.

8. A non-volatile storage medium, characterized in that, The non-volatile storage medium includes a stored program, wherein, when the program is executed, it controls the device containing the non-volatile storage medium to perform the partial discharge signal identification method according to any one of claims 1 to 6.

9. A computer device, characterized in that, include: Memory and processor The memory stores computer programs; The processor is configured to execute a computer program stored in the memory, wherein when the computer program is executed, the processor performs the partial discharge signal identification 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 the processor, it implements the partial discharge signal identification method according to any one of claims 1 to 6.

Images