Method, device and electronic equipment for determining equipment failure results

Abstract

The present invention discloses a method, apparatus, and electronic device for determining equipment failure results. The method comprises: obtaining a discrete waveform sequence of a target device, wherein the discrete waveform sequence includes multiple sampling point data, and the multiple sampling point data include corresponding instantaneous current values ​​at corresponding sampling moments; determining multiple differential waveform sequences based on the discrete waveform sequence; determining probability distributions corresponding to the multiple differential waveform sequences; determining the difference index between any two probability distributions in the multiple probability distributions to obtain multiple difference indices; and determining a failure result of whether the target device is faulty based on the multiple difference indices. The present invention solves the technical problem in related arts that it is difficult to accurately determine whether a device is faulty when the fault signal indicating the fault is relatively weak.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a method and device for determining equipment failure results, and electronic equipment. Background Art

[0002] Power equipment, such as power cables, is a core component for transmitting and distributing electrical energy and is widely used in power generation, transmission, and distribution. Its safe and stable operation is directly related to the reliability of power system power supply and the stability of urban power grids. Compared with traditional overhead lines, power cables offer advantages such as a small footprint, strong anti-interference capabilities, adaptability to complex urban environments, and a low failure rate. They play an irreplaceable role in the construction and upgrade of modern urban power grids. However, due to the long-term underground installation of power cables, their operating environment is complex and volatile, making them susceptible to a variety of factors, including production process defects, accumulated operating time, mechanical stress, moisture, corrosion, and other environmental factors. These factors can cause partial discharge, aging, or defects in the cable insulation, leading to premature failure. Although premature failures have minimal impact on cable operation in the short term, repeated occurrence accelerates insulation degradation, ultimately leading to permanent cable failure, posing a serious threat to the stable operation of the power grid. Effective detection and identification of premature cable failures before they develop into permanent failures, coupled with targeted repair measures, can significantly slow insulation degradation and prevent further power system losses.

[0003] Traditional detection methods suffer from insufficient sensitivity to weak signals in early equipment fault detection. Partial discharge signals are often masked or interfered with by background noise, making it difficult to effectively extract and identify signal features. This lack of sensitivity not only increases the risk of missed detections but also leads to delayed monitoring and early warning of potential faults.

[0004] To address the above-mentioned problems, no effective solutions have been proposed so far. Summary of the Invention

[0005] The embodiments of the present invention provide a method, apparatus, and electronic device for determining a device fault result, so as to at least solve the technical problem in the related art that it is difficult to accurately determine whether a device is faulty when the fault signal indicating the fault is relatively weak.

[0006] According to one aspect of an embodiment of the present invention, a method for determining a device failure result is provided, comprising: obtaining a discrete waveform sequence of a target device, wherein the discrete waveform sequence includes a plurality of sampling point data, and the plurality of sampling point data include corresponding instantaneous values ​​of current at corresponding sampling moments; determining a plurality of differential waveform sequences based on the discrete waveform sequence; determining probability distributions corresponding to the plurality of differential waveform sequences respectively; determining a difference index between any two probability distributions in the plurality of probability distributions to obtain a plurality of difference indexes; and determining a failure result of whether the target device is faulty based on the plurality of difference indexes.

[0007] Optionally, determining the probability distributions corresponding to the multiple differential waveform sequences respectively includes: determining multiple groups of random variable values ​​based on multiple sampling point data respectively included in the multiple differential waveform sequences, wherein the multiple groups of random variable values ​​correspond one-to-one to the multiple differential waveform sequences; determining multiple groups of probability distribution data based on the multiple groups of random variable values, wherein the multiple groups of random variable values ​​correspond one-to-one to the multiple groups of probability distribution data; and determining the probability distributions corresponding to the multiple differential waveform sequences respectively based on the multiple groups of probability distribution data.

[0008] Optionally, determining the probability distributions corresponding to the multiple differential waveform sequences respectively based on the multiple sets of probability distribution data includes: when the probability distributions corresponding to the multiple differential waveform sequences respectively include a first probability distribution and a second probability distribution, determining multiple segment data corresponding to the target probability distribution data, and determining weight data corresponding to the target probability distribution data, wherein the target probability distribution data is any one set of data in the multiple sets of probability distribution data; determining the first probability distribution corresponding to the target probability distribution data based on the multiple segment data, and obtaining the second probability distribution corresponding to the target probability distribution data based on the weight data corresponding to the target probability distribution data.

[0009] Optionally, determining multiple differential waveform sequences based on the discrete waveform sequence includes: determining multiple adjacent periodic sequences based on the discrete waveform sequence; aligning the multiple adjacent periodic sequences to obtain multiple aligned periodic sequences, wherein the reference point data corresponding to each of the multiple aligned periodic sequences are aligned in time, and the multiple adjacent periodic sequences correspond one-to-one to the multiple aligned periodic sequences; and determining the multiple differential waveform sequences based on the multiple aligned periodic sequences.

[0010] Optionally, obtaining the discrete waveform sequence of the target device includes: collecting the initial discrete waveform sequence of the target device; determining filtering parameters related to the target device, wherein the filtering parameters include filtering frequency band parameters; determining impulse response parameters based on the filtering parameters; and filtering the frequency components included in the corresponding current values ​​in the initial discrete waveform sequence based on the impulse response parameters to obtain the discrete waveform sequence.

[0011] Optionally, determining the failure result of whether the target device is faulty based on the multiple difference indices includes: obtaining environmental parameters; determining a predetermined threshold based on the environmental parameters; and determining the failure result of whether the target device is faulty based on the multiple difference indices and the predetermined threshold.

[0012] Optionally, after determining the fault result of whether the target device is faulty based on the multiple difference indexes, the method includes: when the fault result is that the target device is faulty, determining fault parameters based on the multiple difference indexes, wherein the fault parameters include a fault type and a fault level corresponding to the fault type; determining a fault handling strategy corresponding to the fault parameters, and sending the fault parameters to a predetermined terminal.

[0013] According to one aspect of an embodiment of the present invention, a device for determining a device failure result is provided, comprising: an acquisition module for acquiring a discrete waveform sequence of a target device, wherein the discrete waveform sequence includes a plurality of sampling point data, and the plurality of sampling point data include corresponding instantaneous current values ​​at corresponding sampling moments; a first determination module for determining a plurality of differential waveform sequences based on the discrete waveform sequence; a second determination module for determining probability distributions corresponding to the plurality of differential waveform sequences; a third determination module for determining a difference index between any two probability distributions in the plurality of probability distributions to obtain a plurality of difference indexes; and a fourth determination module for determining a failure result of whether the target device is faulty based on the plurality of difference indexes.

[0014] According to one aspect of an embodiment of the present invention, an electronic device is provided, comprising: a processor; and a memory for storing instructions executable by the processor; wherein the processor is configured to execute the instructions to implement any of the above methods for determining a device failure result.

[0015] According to one aspect of an embodiment of the present invention, a computer-readable storage medium is provided. When instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device can execute any of the above-mentioned methods for determining a device failure result.

[0016] In an embodiment of the present invention, a discrete waveform sequence of a target device is obtained, wherein the discrete waveform sequence includes multiple sampling point data, each of which includes instantaneous current values ​​at corresponding sampling moments. Based on the discrete waveform sequence, multiple differential waveform sequences are determined. Probability distributions corresponding to each of the multiple differential waveform sequences are determined. A difference index between any two of the multiple probability distributions is determined to obtain multiple difference indices. Based on the multiple difference indices, a fault result is determined, indicating whether the target device is faulty. This method takes into account the distribution variations between the multiple differential waveform sequences. Because weak fault signals may not be identical in adjacent cycles, traces of the fault are left in the differential waveforms. The corresponding probability distributions can then be used to further amplify the fault or no-fault result. Therefore, multiple differential waveform sequences can be used to accurately determine whether a device is faulty, thereby resolving the technical issue in related arts where it is difficult to accurately determine whether a device is faulty when the fault signal is weak. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0018] Figure 1 is a flow chart of a method for determining a device failure result according to an embodiment of the present invention;

[0019] Figure 2 Schematic diagram of a cable early fault weak signal detection method based on waveform feature point distribution differences provided by an optional embodiment of the present invention;

[0020] Figure 3 4 is a structural block diagram of an apparatus for determining a device failure result according to an embodiment of the present invention. DETAILED DESCRIPTION

[0021] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0022] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0023] Example 1

[0024] According to an embodiment of the present invention, an embodiment of a method for determining the result of a device failure is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0025] Figure 1 FIG. 1 is a flow chart of a method for determining a device failure result according to an embodiment of the present invention. Figure 1 As shown, the method includes the following steps:

[0026] Step S102, obtaining a discrete waveform sequence of the target device, wherein the discrete waveform sequence includes a plurality of sampling point data, and the plurality of sampling point data includes corresponding instantaneous current values ​​at corresponding sampling moments;

[0027] In step S102 provided in the present application, a discrete waveform sequence of the target device is obtained.

[0028] This involves the discrete waveform sequence of the target device, a sequence of discrete data points obtained by sampling the current signal of the target device (such as a cable). In power systems, the operating status of a device can be reflected by its current or voltage waveform. The discrete waveform sequence converts this continuous waveform signal into a series of instantaneous values ​​distributed discretely in time, facilitating storage and digital signal processing.

[0029] This involves multiple sampling points, meaning that the signal sampler regularly collects instantaneous signal values ​​at multiple points in time within a certain time interval. The sampling frequency can be adaptively set based on the actual application and scenario. For example, a higher sampling frequency can capture more details. High-frequency components, such as those of partial discharge, are more representative of the characteristic fault signals.

[0030] This involves the instantaneous current value at the corresponding sampling moment, that is, the signal strength measured by the sampler at each sampling moment. In power systems, the instantaneous current value directly reflects the operating status of equipment. By detecting changes in the current waveform, the difference between normal and abnormal equipment operation can be identified.

[0031] Step S104, determining a plurality of differential waveform sequences based on the discrete waveform sequence;

[0032] In step S104 provided in the present application, a plurality of differential waveform sequences are determined.

[0033] This involves differential waveform sequences, which are new signal sequences derived by calculating the difference between the instantaneous values ​​of corresponding sampling points in two adjacent discrete waveform sequences. The purpose of calculating differential waveform sequences is generally to highlight the dynamic characteristics of signal changes and eliminate or suppress the influence of trending background signals, thereby making it easier to identify abnormalities or fault characteristics in the signal. In the early detection of power cable faults, differential waveform sequences can enhance weak fault signals, making them more prominent amidst complex background noise.

[0034] By calculating the differential waveform sequence, signal changes can be highlighted, helping to identify subtle changes that might be masked by background noise or trend signals in the original waveform sequence, thereby improving signal detection sensitivity. Furthermore, the differential waveform sequence can suppress static or trend noise in the signal, as such noise varies little between adjacent cycles. True fault-related signal changes are amplified in the differential sequence, thus achieving signal-to-noise separation. For weak signals indicating early-stage power cable faults, the differential waveform sequence can capture their dynamic characteristics, enabling identification even of very weak signals. This significantly improves the accuracy and timeliness of early-stage fault detection.

[0035] Step S106, determining probability distributions corresponding to the plurality of differential waveform sequences;

[0036] In step S106 provided in the present application, probability distributions corresponding to the plurality of differential waveform sequences are determined.

[0037] This involves probability distribution, which is a set of functions or graphs that describe the possible values ​​of a random variable and their probabilities. In signal processing, probability distributions can reflect the statistical characteristics of the signal value range, such as the frequency, mean, and variance of the signal value. Constructing the probability distribution of multiple differential waveform sequences means statistically analyzing the differences between each sampling point in these sequences to determine the probability of these differences occurring within the entire value range. In the power cable inspection process, each differential waveform sequence corresponds to a probability distribution, which reflects the statistical characteristics of the cable current waveform at different time points. These probability distributions may vary depending on the device status, operating conditions, or the presence of potential faults.

[0038] The construction of probability distribution allows the use of different statistical methods for fault detection, such as histogram analysis, kernel density estimation, relative entropy KL divergence calculation, etc., which increases the diversity and flexibility of fault detection strategies.

[0039] Step S108, determining a difference index between any two probability distributions in the plurality of probability distributions to obtain a plurality of difference indexes;

[0040] In step S108 provided in the present application, a difference index between any two probability distributions in a plurality of probability distributions is determined, until a difference index between every two probability distributions in a plurality of probability distributions is determined, thereby obtaining a plurality of difference indices.

[0041] This involves the difference index, a quantitative metric used to measure the statistical difference between two probability distributions. In signal processing and data analysis, difference indices can include the Kullback-Leibler divergence (KL divergence), Jensen-Shannon divergence, Bhattacharyya distance, or Wasserstein distance. These difference indices help assess the similarity or distance between the current signal state and a known normal state.

[0042] For probability distributions constructed from multiple differential waveform sequences, the difference index between any two distributions is calculated, resulting in a series of difference index values. This step essentially evaluates changes in the signal distribution at different time points or under different operating conditions of the cable, specifically looking for time points where significant differences from the normal operating state distribution occur, thereby identifying signs of potential early faults.

[0043] Step S110 , determining whether the target device is faulty based on the plurality of difference indices.

[0044] In step S110 provided in this application, the final failure result is determined.

[0045] Among them, the fault result is involved. The fault result refers to the difference index obtained through analysis and calculation to determine whether the equipment, such as the cable, has an early fault or abnormal state.

[0046] Multiple difference indices allow for early detection of minor anomalies in power cables, even if they are very subtle at an early stage, helping to take corrective measures before faults develop into more serious problems.

[0047] Through steps S102-S110, a discrete waveform sequence of the target device is obtained. The discrete waveform sequence includes multiple sampling point data, each of which includes instantaneous current values ​​at corresponding sampling moments. Based on the discrete waveform sequence, multiple differential waveform sequences are determined. Probability distributions corresponding to each of the multiple differential waveform sequences are determined. Difference indices between any two of the multiple probability distributions are determined to obtain multiple difference indices. Based on the multiple difference indices, a fault result is determined, indicating whether the target device is faulty. This method takes into account the distribution variations between the multiple differential waveform sequences. Because weak fault signals may not be identical in adjacent cycles, traces of the fault are left in the differential waveforms. The corresponding probability distributions can then be used to further amplify the fault or no-fault result. Therefore, multiple differential waveform sequences can be used to accurately determine whether the device is faulty, thereby resolving the technical issue in related art where it is difficult to accurately determine whether the device is faulty when the fault signal is weak.

[0048] As an optional embodiment, determining the probability distributions corresponding to multiple differential waveform sequences respectively includes: determining multiple groups of random variable values ​​based on multiple sampling point data respectively included in the multiple differential waveform sequences, wherein the multiple groups of random variable values ​​correspond one-to-one to the multiple differential waveform sequences; determining multiple groups of probability distribution data based on the multiple groups of random variable values, wherein the multiple groups of random variable values ​​correspond one-to-one to the multiple groups of probability distribution data; and determining the probability distributions corresponding to the multiple differential waveform sequences respectively based on the multiple groups of probability distribution data.

[0049] In this embodiment, a process of constructing a probability distribution from a differential waveform sequence collected from a power cable is described.

[0050] This involves multiple sets of random variable values. In statistics, a random variable is a mathematical model used to describe uncertain events. Multiple sets of random variable values ​​refer to treating the data at each sampling point in the differential waveform sequence as the value of a random variable. Due to the uncertainty of current changes during cable operation, these differential values ​​can be treated as random variables, and their statistical properties (such as mean and variance) reflect the fluctuations in cable conditions.

[0051] This involves multiple sets of probability distribution data, which describes the possible values ​​within the range of a random variable and their frequency of occurrence. In power cable testing, this typically means converting each differential waveform sequence into a probability distribution over its range, which helps identify anomalous signal patterns.

[0052] This involves constructing a probability distribution corresponding to each of the multiple differential waveform sequences. This means constructing a probability distribution for each differential waveform sequence, which describes the statistical characteristics of the sample point values ​​in the sequence. By comparing these probability distributions with the probability distribution under known normal conditions, the degree of variation in the differential waveform sequence can be quantified.

[0053] In this step, the sampling point data in a series of differential waveform sequences is treated as random variables. The probability distribution data for these random variables is then constructed, ultimately yielding a probability distribution corresponding to each sequence. This process converts the time-domain signal into a statistical representation, helping to identify weak abnormal signals in a noisy background.

[0054] Probability distributions can highlight the statistical characteristics of signals, such as mean and variance, which are crucial for extracting subtle changes in early-stage fault signals. They can also improve the sensitivity of fault detection. By using changes in probability distributions, potential faults can be identified even with very small signal amplitudes, thus increasing the sensitivity of fault detection. Furthermore, the construction of probability distributions is often accompanied by noise suppression, which makes the detection method more robust to background noise and more accurately identifies true signal changes.

[0055] As an optional embodiment, based on multiple groups of probability distribution data, probability distributions corresponding to multiple differential waveform sequences are determined, including: when the probability distributions corresponding to the multiple differential waveform sequences include a first probability distribution and a second probability distribution, multiple segment data corresponding to the target probability distribution data are determined, and weight data corresponding to the target probability distribution data is determined, wherein the target probability distribution data is any group of data in the multiple groups of probability distribution data; based on the multiple segment data, a first probability distribution corresponding to the target probability distribution data is determined, and based on the weight data corresponding to the target probability distribution data, a second probability distribution corresponding to the target probability distribution data is obtained.

[0056] In this embodiment, it is described how to determine a plurality of probability distributions corresponding to a specific differential waveform sequence from a plurality of probability distribution data.

[0057] This involves a first probability distribution and a second probability distribution.

[0058] The first probability distribution can be calculated in a discrete form. Discrete form calculations construct a discrete probability distribution by dividing the data range into multiple bins and counting the frequency of data occurrences within each bin. Each bin represents a range of data values, and the frequency indicates the number of times the data occurs within that range. After normalizing the frequencies, the probability of data occurrence in each bin is obtained. In discrete form calculations, although we process data from all sampling points, we calculate probabilities based on bins rather than individual sampling points. This simplifies calculations and facilitates processing and comparison, but it may lose some details about the specific distribution shape of the data.

[0059] The second probability distribution can be a distribution calculated in the form of kernel density estimation. Kernel density estimation (KDE) is a non-parametric probability density function (PDF) estimation method that attempts to smoothly estimate a continuous PDF from the data. In kernel density estimation, each sampling point is actually assigned a weight, which is calculated by a kernel function (such as a Gaussian kernel), and the size of the weight depends on the distance between the sampling point and the observation point and the bandwidth parameter of the kernel function. KDE obtains the PDF estimate at the observation point by summing the weights of all sampling points and normalizing them. In this way, KDE takes into account all sampling point data, but unlike discrete form calculations, it can provide a continuous and smooth probability density function estimate, retaining more details of the data distribution, and is especially flexible and accurate when estimating complex distribution shapes.

[0060] This involves target probability distribution data and weight data. The target probability distribution data is any one of multiple probability distribution data sets, typically corresponding to the probability distribution of the most recently acquired differential waveform sequence. Weight data refers to the relative importance assigned to different data points or distributions when fusing or adjusting probability distributions. It can be dynamically adjusted based on actual conditions (such as signal strength, noise level, and device status).

[0061] In this way, both the first probability distribution and the second probability distribution are taken into consideration, and the fault result is determined by integrating the two aspects, thereby enhancing the accuracy of the fault result determination.

[0062] As an optional embodiment, multiple differential waveform sequences are determined based on the discrete waveform sequence, including: determining multiple adjacent periodic sequences based on the discrete waveform sequence; aligning the multiple adjacent periodic sequences to obtain multiple aligned periodic sequences, wherein the reference point data corresponding to the multiple aligned periodic sequences are aligned in time, and the multiple adjacent periodic sequences correspond one-to-one to the multiple aligned periodic sequences; and determining multiple differential waveform sequences based on the multiple aligned periodic sequences.

[0063] In this embodiment, a detailed process of generating multiple differential waveform sequences from an originally acquired discrete waveform sequence in early fault detection of a power cable is described.

[0064] This involves adjacent cycle sequences, which refer to discrete waveform sequences within two consecutive power frequency cycles. Since power cable signals are typically periodic, especially power frequency current signals, comparing signal changes within these cycles can capture subtle fault signals or changes in device status.

[0065] This involves alignment, which involves performing temporal alignment between adjacent cycle sequences to ensure synchronization of signals within multiple cycles, thereby preventing phase differences from affecting the accuracy of subsequent differential waveform sequences. In power signal processing, this is typically achieved by detecting zero crossings, peaks, or other synchronization markers.

[0066] This involves multiple aligned cycle sequences and multiple differential waveform sequences. After alignment, multiple differential waveform sequences are generated by performing point-by-point differential operations on the sampled values ​​of each aligned cycle sequence and the next. These sequences highlight inter-cycle variations in the signal and are particularly effective for detecting small fluctuations in the early stages of faults.

[0067] From the original discrete waveform sequence, we first select multiple adjacent periodic sequences containing consecutive power frequency cycles. To ensure that these periodic sequences are strictly aligned in time, we align them to obtain multiple aligned periodic sequences. By performing a differential operation on the aligned sequences, we ultimately generate multiple differential waveform sequences. These sequences highlight the inter-cycle variations in cable signals and serve as sensitive indicators for early fault detection.

[0068] Differential waveform sequences can enhance weak signals, particularly those associated with incipient faults, as these signals may exhibit inconsistent patterns within adjacent cycles. Alignment processing can eliminate phase differences in periodic signals, thereby suppressing the effects of periodic noise and improving fault signal recognition. Differential waveform sequences can intuitively quantify the degree of signal variation between cycles, making them particularly useful for analyzing fluctuations in cable operating conditions.

[0069] As an optional embodiment, obtaining a discrete waveform sequence of a target device includes: collecting an initial discrete waveform sequence of the target device; determining filtering parameters corresponding to the target device, wherein the filtering parameters include filtering frequency band parameters; determining impulse response parameters based on the filtering parameters; and filtering the frequency components included in the corresponding current values ​​in the initial discrete waveform sequence based on the impulse response parameters to obtain a discrete waveform sequence.

[0070] In this embodiment, it is described how to extract a clear waveform sequence from the collected original signal in the early stage fault detection of the power cable so as to perform subsequent fault analysis.

[0071] This involves filtering parameters, which are a set of parameters used to set filter characteristics, including filter band parameters (such as passband frequency and stopband frequency). Filter band parameters define the frequency range that the filter allows to pass. For power signal processing, this usually involves removing or suppressing unwanted frequency components, such as noise or interference signals.

[0072] This involves impulse response parameters. The impulse response is the filter's response to a single impulse signal, which describes the filter's characteristics. Impulse response parameters include filter type (low-pass, high-pass, band-pass, etc.), cutoff frequency, and filter order.

[0073] This involves filtering the initial discrete waveform sequence. This involves applying a filter to the original discrete waveform sequence to remove noise and unwanted frequency components while retaining the signal relevant to the analysis target. In this application, the goal is to retain signal components that can indicate early-stage faults in power cables, such as those caused by partial discharges or subtle changes in insulation materials.

[0074] The raw discrete waveform sequence acquired from the target device may contain significant noise and unwanted frequency components, which can mask weak signals of incipient faults. Therefore, the filtering parameters, including the filter frequency band, are first determined to define the filter's passband and stopband. Then, based on these filtering parameters, the filter's impulse response parameters are determined, and a filter model suitable for power cable signal processing is constructed. Finally, this filter is applied to the raw discrete waveform sequence, retaining signal components associated with incipient faults. This results in a clearer and more accurate discrete waveform sequence, providing a clean signal input for subsequent fault detection and analysis.

[0075] As an optional embodiment, determining whether the target device is faulty based on multiple difference indexes includes: obtaining environmental parameters; determining a predetermined threshold based on the environmental parameters; and determining whether the target device is faulty based on multiple difference indexes and the predetermined threshold.

[0076] In this embodiment, a specific process of combining environmental parameters and difference indexes to determine whether a target device (such as a power cable) has a fault in early power cable fault detection is described.

[0077] Environmental parameters are considered as external factors that affect the operation of power cables, including but not limited to temperature, humidity, cable load, and electromagnetic interference levels. These parameters significantly affect the performance and signal characteristics of the cable and need to be considered when determining faults.

[0078] This involves a predetermined threshold, which is pre-set in the fault detection algorithm to determine whether the difference index reaches the threshold value of the fault warning standard. This threshold is usually determined based on historical data analysis, device characteristics, and the influence of environmental parameters. It can help distinguish normal fluctuations from fault signals.

[0079] In power cable fault detection, multiple difference indices are first calculated. These indices quantify the differences between the statistical distribution characteristics of the cable current waveform at different time points and its normal state. Next, the current environmental parameters are obtained. This is because environmental factors such as temperature, humidity, and electromagnetic interference can affect the value of the difference indices, potentially affecting the accuracy of fault diagnosis. Next, based on the influence of environmental parameters, a predetermined threshold is dynamically determined to adapt to the current operating conditions. Finally, by comparing the multiple difference indices with the adjusted predetermined threshold, a determination is made as to whether the target device has experienced a fault. This process ensures the accuracy and reliability of fault detection and is adaptable to various operating environments and changes in device status. By dynamically adjusting the predetermined threshold based on environmental parameters, the fault detection system can adapt to various operating conditions and improve detection sensitivity and accuracy.

[0080] As an optional embodiment, after determining the fault result of whether the target device is faulty based on multiple difference indexes, it includes: when the fault result is that the target device is faulty, determining fault parameters based on multiple difference indexes, wherein the fault parameters include the fault type and the fault level corresponding to the fault type; determining a fault handling strategy corresponding to the fault parameters, and sending the fault parameters to a predetermined terminal.

[0081] This embodiment describes how to further analyze the fault characteristics and adopt targeted processing strategies after confirming the equipment fault in the early stage fault detection of the power cable.

[0082] This involves fault parameters, which are detailed information used to classify and grade detected faults. These parameters include the fault type (such as insulation aging, partial discharge, conductor breakage, etc.) and the corresponding fault severity (such as minor, moderate, severe, etc.). These parameters help to more accurately identify the nature and severity of the fault.

[0083] This involves fault handling strategies, which are specific response measures designed based on fault parameters. These include, but are not limited to, monitoring upgrades, preventive maintenance, fault location, and emergency shutdowns. These strategies should ensure the safety and stability of the power system while considering the economic cost of the fault and its impact on grid operations.

[0084] This involves a predetermined terminal, which is the destination for receiving fault parameters and handling strategy information. This can be a computer system in a monitoring center, a maintenance personnel's mobile device, an automated control device, etc. By sending fault parameters to the predetermined terminal, a rapid response and handling of the fault can be ensured.

[0085] By analyzing multiple difference indices, fault types and locations can be more accurately identified, providing a foundation for targeted troubleshooting. Setting fault levels helps maintenance personnel prioritize faults based on severity, avoiding resource waste and improving system maintenance efficiency. Directly transmitting fault parameter information to a predetermined terminal enables real-time transmission of fault information, ensuring rapid response and resolution, and minimizing the impact of faults on the power grid. Fault handling strategies are developed based on specific fault parameters, helping to reduce operational errors and improve the operational safety of the power system.

[0086] Based on the above embodiment and optional embodiment, an optional implementation manner is provided, which is described in detail below.

[0087] This optional embodiment of the present invention proposes a method for early fault detection based on the deviation between adjacent cycle differential waveforms and a Gaussian distribution, targeting weak current disturbance signals that may occur in power equipment, particularly cables, during early fault conditions. By collecting data from multi-cycle power frequency current waveforms, constructing differential waveforms, extracting distribution features, and calculating KL divergence, combined with threshold determination, this method achieves highly sensitive detection and early warning of weak signals from early cable faults. Figure 2 This is a schematic diagram of a cable early fault weak signal detection method based on waveform feature point distribution differences provided by an optional embodiment of the present invention. Figure 2 Introduce it:

[0088] The current waveform is more sensitive to the condition of cable equipment and can be used to detect weak early-stage fault signals. Using the instantaneous change in the current waveform as a characteristic, the system takes the difference between adjacent cycle waveform signals and uses the distribution of the signal difference as a typical characteristic. Anomalies are detected by comparing the statistical distribution of waveform changes with and without early-stage fault disturbances. If the equipment is operating normally, the differential waveform contains only noise information and follows a Gaussian statistical distribution. If the equipment is experiencing an early-stage fault, the differential waveform contains both noise and fault signals. Therefore, the statistical distribution of the differential waveform containing fault information deviates from a Gaussian distribution. Therefore, the KL divergence (KLD) can be used to evaluate the KLD distance between the differential waveform in the early-stage fault state and the Gaussian distribution of the normal differential waveform. Based on a set threshold, the presence of a fault can be determined. If the KLD exceeds the threshold, an early-stage fault disturbance is present.

[0089] The method comprises the following steps:

[0090] 1. Data collection and preprocessing

[0091] (1) Data collection

[0092] In normal operation or test environment of the cable, use high-speed data acquisition device to collect the power frequency current waveform. Assume that the power frequency is (such as 50Hz), select the sampling rate (e.g. 10kHz, 50kHz or higher) in order to capture the high-frequency components that may be contained in early faults. If the number of sampling points collected in each power frequency cycle is:

[0093]

[0094] Collect several consecutive cycles (denoted as The total number of sampling points is approximately In high voltage or ultra-high voltage environments, optical current sensors or high-precision Hall current sensors can be used to ensure high bandwidth and high resolution of signal measurement.

[0095] (2) Data preprocessing

[0096] In order to avoid the masking of early fault signals by the power frequency fundamental wave and environmental noise, the collected discrete waveform sequence is usually Perform bandpass filtering. If the impulse response of the filter in the digital domain is And the length is , it can be expressed as:

[0097]

[0098] Get the filtered signal The filtering range should cover the potential fault frequency band (such as hundreds of Hz to tens of kHz) while suppressing both very low-frequency and very high-frequency noise. h[n] represents the impulse response coefficient of the filter at time point n, while I[kn] represents the value of the input signal I at time kn.

[0099] 2. Construction of differential waveforms between adjacent cycles

[0100] (1) Periodic registration

[0101] There may be a slight phase shift between adjacent power frequency cycles of the current waveform. By using the synchronization signal or zero-crossing point alignment method, it can be ensured that the starting point (or any point) of each cycle is strictly aligned in time. Cycle includes Sampling points:

[0102]

[0103] Here, M represents the number of groups or samples in the data set or signal sequence. Specifically in this scenario, it represents the number of groups of differential waveforms.

[0104] (2) Differential operation

[0105] After completing the cycle registration, the adjacent cycles are subtracted point by point to obtain the differential waveform sequence. Cycle and The sampling values ​​of the period are recorded as:

[0106]

[0107] Wherein, k is an index used to indicate the specific sample point position in the differential waveform sequence.

[0108] The differential waveform Defined as:

[0109]

[0110] When a collection When there are consecutive cycles, we can get Group differential waveform Differential processing can suppress the power frequency fundamental wave and conventional noise, making the subtle anomalies caused by early faults stand out.

[0111] 3. Differential waveform distribution feature extraction

[0112] (1) Probability density estimation

[0113] The differential waveform All sampling points is considered as the value of a random variable.

[0114] To quantify its statistical laws, a histogram can be constructed or kernel density estimation (KDE) can be used to approximate the probability density function.

[0115] 1) Histogram method

[0116] The range of the difference value Divided equally interval, the The probability of an interval is defined as:

[0117]

[0118] in, Indicates falling into The number of difference samples in the interval, is the total number of sampling points of the differential waveform.

[0119] 2) Kernel Density Estimation (KDE) For a Gaussian kernel, the estimated probability density function can be written as:

[0120]

[0121] in, is the bandwidth parameter, is the sampling value of the differential waveform, and x is the value of the continuous variable, representing the point whose probability density is to be estimated. KDE can produce a smoother distribution curve.

[0122] (2) Establishment of reference distribution

[0123] When the cable is generally considered normal, multiple sets of differential waveforms are collected and their overall probability distribution is estimated, which is recorded as If the distribution is close to Gaussian, is the standard deviation, is the mean, is the variance, that is:

[0124]

[0125] If there is a deviation from the Gaussian assumption, the measured distribution can be directly saved as a discrete form or kernel density function and used as a benchmark for subsequent comparisons.

[0126] 4. KLD (KL divergence) calculation

[0127] (1) KLD concept

[0128] KLD is an important indicator for evaluating the difference between two probability distributions. Assume that the reference distribution is , the real-time measured distribution is , then in continuous form we have:

[0129]

[0130] when and Exactly the same, then The greater the difference between the two, The larger the value.

[0131] (2) Discretization implementation

[0132] In practice, KLD is often calculated by using the discrete probability of histogram or kernel density estimation. If the histogram method is used to discretize the differential waveform distribution into , the reference distribution is , then KLD is approximately:

[0133]

[0134] To avoid or The resulting numerical divergence can be used to smooth the distribution probability or set a minimum threshold (like ) instead of a zero value.

[0135] 5. Threshold determination

[0136] (1) Threshold setting

[0137] In the normal state of the cable without faults, the KLD threshold can be determined by statistically analyzing the KLD value distribution of a large number of differential waveforms, combined with a certain confidence interval (such as 95% or 99%) or field test fault simulation results. When the noise level or load conditions change significantly, an adaptive threshold strategy can also be used to improve the reliability of detection.

[0138] (2) Fault diagnosis

[0139] Let the divergence value calculated in real time be:

[0140]

[0141] If the cable has an early fault disturbance, it is judged that the cable has an early fault disturbance; otherwise, the device is considered to be in normal state. For edge cases, a threshold interval can be set , strengthen subsequent data collection and observation.

[0142] (3) Multiple confirmation mechanism

[0143] In order to reduce misjudgment caused by transient spikes or one-time interference, multiple sets of differential waveforms can be continuously acquired within a short period of time (such as 0.5 seconds or 1 second) and calculated separately.

[0144] If the threshold is exceeded continuously for at least several times, a fault alarm will be issued. If the threshold is exceeded only once or accidentally, it will be temporarily recorded as a suspicious event and monitoring will continue.

[0145] 6. Fault output and early warning

[0146] (1) Alarm and data recording

[0147] Once the fault state is determined, the system immediately triggers an alarm and sends the current differential waveform, sampling time, noise level, load information and The value is stored in the database and the alarm information is sent to the dispatch center or maintenance personnel.

[0148] (2) Early warning classification and subsequent evaluation

[0149] according to With threshold The fault risk can be divided into different levels based on the gap between the alarms. If the alarms appear continuously and the trend becomes more obvious, other diagnostic methods (such as partial discharge measurement and harmonic energy analysis) can be combined to locate the fault area and arrange maintenance as soon as possible.

[0150] Through the above optional implementation, at least the following beneficial effects can be achieved:

[0151] (1) The present invention uses KL divergence to quantitatively analyze the difference between the differential distribution of current waveforms and the normal Gaussian distribution, and can accurately capture subtle changes in early fault signals. This method can not only effectively distinguish between noise in normal conditions and fault signals in fault conditions, but also significantly improve the sensitivity and accuracy of detection. By taking the difference between the current waveforms of adjacent cycles, extracting statistical features, and comparing them with the Gaussian distribution, it ensures that anomalies can be detected in the early stages of a fault, reducing the probability of missed detection and false detection.

[0152] (2) The method of the present invention is based on KL divergence analysis of waveform differential distribution and has strong noise resistance. In actual power systems, current waveforms are often interfered with by various noises, such as electromagnetic interference and environmental noise. Traditional time domain and frequency domain analysis methods are easily masked by noise in high-noise environments, resulting in reduced detection performance. However, the present invention can effectively suppress the influence of noise by quantifying the statistical distribution differences, ensuring that fault signals can still be accurately detected in complex noise backgrounds.

[0153] (3) By analyzing the statistical characteristics of the differential distribution of current waveforms, the method of the present invention has the ability to adapt to a variety of different types of faults, such as insulation aging, partial discharge, and conductor breakage. Because different fault types have different effects on the current waveform, this method does not rely on specific fault characteristics. Instead, it quantifies the overall distribution differences through KL divergence, thereby achieving unified detection of multiple faults. This versatility enables the method to maintain good detection performance in the face of diverse faults, improving the adaptability and flexibility of the system.

[0154] It should be noted that for the aforementioned method embodiments, for simplicity of description, they are all expressed as a series of action combinations. However, those skilled in the art should be aware that the present invention is not limited by the order of the actions described, because according to the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in this specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0155] Through the description of the above embodiments, those skilled in the art will clearly understand that the methods according to the above embodiments can be implemented using software plus the necessary general-purpose hardware platform. Of course, hardware can also be used, but in many cases the former is the more preferred embodiment. Based on this understanding, the technical solution of the present invention, or the portion 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, or optical disk) and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0156] Example 2

[0157] According to an embodiment of the present invention, there is also provided a device for implementing the above-mentioned method for determining the result of a device failure. Figure 3 FIG. 1 is a structural block diagram of an apparatus for determining a device failure result according to an embodiment of the present invention. Figure 3 As shown, the apparatus includes: an acquisition module 302, a first determination module 304, a second determination module 306, a third determination module 308 and a fourth determination module 310. The apparatus will be described in detail below.

[0158] An acquisition module 302 is configured to acquire a discrete waveform sequence of a target device, wherein the discrete waveform sequence includes data at multiple sampling points, and the data at multiple sampling points include instantaneous current values ​​corresponding to corresponding sampling moments. A first determination module 304 is connected to the acquisition module 302 and configured to determine multiple differential waveform sequences based on the discrete waveform sequence. A second determination module 306 is connected to the first determination module 304 and configured to determine probability distributions corresponding to the multiple differential waveform sequences. A third determination module 308 is connected to the second determination module 306 and configured to determine a difference index between any two probability distributions in the multiple probability distributions to obtain multiple difference indexes. A fourth determination module 310 is connected to the third determination module 308 and configured to determine a fault result indicating whether the target device is faulty based on the multiple difference indexes.

[0159] It should be noted here that the above-mentioned acquisition module 302, first determination module 304, second determination module 306, third determination module 308 and fourth determination module 310 correspond to steps S102 to S110 in the method for determining the results of equipment failure. The instances and application scenarios implemented by multiple modules and corresponding steps are the same, but are not limited to the contents disclosed in the above-mentioned embodiment 1.

[0160] Example 3

[0161] According to another aspect of an embodiment of the present invention, an electronic device is provided, including: a processor; and a memory for storing processor-executable instructions, wherein the processor is configured to execute the instructions to implement any of the above methods for determining device failure results.

[0162] Example 4

[0163] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided. When the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device can execute any of the above-mentioned methods for determining device failure results.

[0164] The serial numbers of the above embodiments of the present invention are for description only and do not represent the advantages or disadvantages of the embodiments.

[0165] In the above embodiments of the present invention, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0166] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only exemplary. For example, the division of the units can be a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0167] The units described as separate components may or may not be physically separate, and 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 may be selected according to actual needs to achieve the purpose of the present embodiment.

[0168] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0169] 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 computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion 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 for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), a mobile hard drive, a magnetic disk, or an optical disk.

[0170] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A method for determining the result of a device failure, characterized in that: include: Acquire a discrete waveform sequence of a target device, wherein the discrete waveform sequence includes a plurality of sampling point data, and the plurality of sampling point data includes corresponding instantaneous current values ​​at corresponding sampling moments; Determining a plurality of differential waveform sequences based on the discrete waveform sequence; Determining probability distributions corresponding to the plurality of differential waveform sequences respectively; Determine a difference index between any two probability distributions among the plurality of probability distributions to obtain a plurality of difference indices; Determining a failure result of whether the target device is faulty based on the multiple difference indexes; Wherein, determining the probability distributions corresponding to the plurality of differential waveform sequences respectively includes: determining a plurality of groups of random variable values ​​based on a plurality of sampling point data respectively included in the plurality of differential waveform sequences, wherein the plurality of groups of random variable values ​​correspond one-to-one to the plurality of differential waveform sequences; determining a plurality of groups of probability distribution data based on the plurality of groups of random variable values, wherein the plurality of groups of random variable values ​​correspond one-to-one to the plurality of groups of probability distribution data; and determining the probability distributions corresponding to the plurality of differential waveform sequences respectively based on the plurality of groups of probability distribution data; Wherein, determining the probability distributions corresponding to the multiple differential waveform sequences respectively based on the multiple sets of probability distribution data includes: when the probability distributions corresponding to the multiple differential waveform sequences respectively include a first probability distribution and a second probability distribution, determining multiple segment data corresponding to the target probability distribution data, and determining weight data corresponding to the target probability distribution data, wherein the target probability distribution data is any one set of data in the multiple sets of probability distribution data; determining the first probability distribution corresponding to the target probability distribution data based on the multiple segment data, and obtaining the second probability distribution corresponding to the target probability distribution data based on the weight data corresponding to the target probability distribution data, wherein the first probability distribution is a distribution in a discrete form, and the second probability distribution is a distribution calculated in the form of kernel density estimation.

2. The method according to claim 1, characterized in that Determining a plurality of differential waveform sequences based on the discrete waveform sequence includes: Determining a plurality of adjacent period sequences based on the discrete waveform sequence; Performing alignment processing on the multiple adjacent periodic sequences to obtain multiple aligned periodic sequences, wherein the reference point data respectively corresponding to the multiple aligned periodic sequences are aligned in time, and the multiple adjacent periodic sequences correspond to the multiple aligned periodic sequences in a one-to-one manner; The plurality of differential waveform sequences are determined according to the plurality of aligned period sequences.

3. The method according to claim 1, characterized in that The step of obtaining a discrete waveform sequence of a target device includes: Acquiring an initial discrete waveform sequence of the target device; Determining filtering parameters for the target device, wherein the filtering parameters include filtering frequency band parameters; Determining impulse response parameters based on the filtering parameters; The frequency components included in the current values ​​corresponding to the initial discrete waveform sequence are filtered according to the impulse response parameters to obtain the discrete waveform sequence.

4. The method according to claim 1, wherein Determining, based on the plurality of difference indexes, whether the target device is faulty, includes: Get environmental parameters; determining a predetermined threshold value based on the environmental parameter; A failure result is determined as to whether the target device is faulty based on the multiple difference indexes and the predetermined threshold.

5. The method according to any one of claims 1 to 4, characterized in that After determining whether the target device is faulty based on the plurality of difference indices, the method further includes: In a case where the fault result is a fault of the target device, determining a fault parameter according to the multiple difference indexes, wherein the fault parameter includes a fault type and a fault level corresponding to the fault type; A fault processing strategy corresponding to the fault parameters is determined, and the fault parameters are sent to a predetermined terminal.

6. A device for determining equipment failure results, characterized in that: include: An acquisition module, configured to acquire a discrete waveform sequence of a target device, wherein the discrete waveform sequence includes a plurality of sampling point data, and the plurality of sampling point data includes corresponding instantaneous current values ​​at corresponding sampling moments; A first determining module is configured to determine a plurality of differential waveform sequences based on the discrete waveform sequence; A second determining module is used to determine the probability distributions corresponding to the plurality of differential waveform sequences respectively; A third determination module is used to determine a difference index between any two probability distributions in the plurality of probability distributions to obtain a plurality of difference indexes; a fourth determining module, configured to determine a failure result of whether the target device is faulty based on the multiple difference indexes; The second determining module is further configured to determine, based on the plurality of sampling point data respectively included in the plurality of differential waveform sequences, a plurality of groups of random variable values, wherein the plurality of groups of random variable values ​​correspond one-to-one to the plurality of differential waveform sequences; determine, based on the plurality of groups of random variable values, a plurality of groups of probability distribution data, wherein the plurality of groups of random variable values ​​correspond one-to-one to the plurality of groups of probability distribution data; and determine, based on the plurality of groups of probability distribution data, probability distributions corresponding to the plurality of differential waveform sequences respectively; Wherein, the second determination module is further used to determine the multi-segment data corresponding to the target probability distribution data and determine the weight data corresponding to the target probability distribution data when the probability distributions corresponding to the multiple differential waveform sequences respectively include the first probability distribution and the second probability distribution, wherein the target probability distribution data is any one group of data in the multiple groups of probability distribution data; based on the multi-segment data, determine the first probability distribution corresponding to the target probability distribution data, and based on the weight data corresponding to the target probability distribution data, obtain the second probability distribution corresponding to the target probability distribution data, wherein the first probability distribution is a distribution in a discrete form, and the second probability distribution is a distribution calculated in the form of kernel density estimation.

7. An electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is configured to execute the instructions to implement the method for determining a device failure result according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that When the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the method for determining a device failure result according to any one of claims 1 to 5.

Images