Signal type determination method and device and electronic equipment

Abstract

The invention discloses a signal type determination method and device and electronic equipment. The method comprises the following steps: acquiring a signal to be detected; determining a target signal feature corresponding to the to-be-detected signal; calling a clustering space corresponding to the signal parameters of the to-be-detected signal; obtaining a plurality of feature distances according to the target signal features and the plurality of clustering centers; and according to the plurality of feature distances, determining a target signal type corresponding to the to-be-detected signal from the plurality of to-be-selected signal types. According to the method and the device, the technical problem that the accuracy of the determined signal type is low when multiple discharge types of signals or interference signals exist in the detection environment in the prior art is solved.

Description

Technical Field

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

[0002] Determining the signal type of a partial discharge signal is of great significance for designing an overhaul strategy for high-voltage cables and ensuring the safe and reliable operation of a power transmission system. Currently, the phase pattern recognition method is mainly used to determine the signal type of a partial discharge signal. However, when there are signals of multiple discharge types or interference signals in the detection environment, the accuracy of identifying the signal type based on the signal pattern is greatly reduced.

[0003] In view of the above problems, no effective solution has been proposed yet. Summary of the Invention

[0004] Embodiments of the present invention provide a method, apparatus, and electronic device for determining a signal type, so as to at least solve the technical problem of low accuracy of the determined signal type when there are signals of multiple discharge types or interference signals in the detection environment in the related art.

[0005] According to an aspect of an embodiment of the present invention, there is provided a method for determining a signal type, including: obtaining a signal to be detected, where a waveform change frequency value corresponding to the signal to be detected is greater than a waveform change frequency threshold; determining a target signal feature corresponding to the signal to be detected; retrieving a clustering space corresponding to signal parameters of the signal to be detected, where the clustering space includes a plurality of cluster centers, and the plurality of cluster centers correspond one-to-one to a plurality of candidate signal types, and the plurality of candidate signal types are signal types for which the probability of belonging to the signal type corresponding to the signal parameters is greater than a predetermined threshold; obtaining a plurality of feature distances based on the target signal feature and the plurality of cluster centers; and determining a target signal type corresponding to the signal to be detected from the plurality of candidate signal types based on the plurality of feature distances.

[0006] Optionally, before obtaining the signal to be detected, the method further includes: obtaining an original signal; determining, based on the original signal, a plurality of segmented signals corresponding to a plurality of predetermined time periods respectively, where the plurality of predetermined time periods are continuous on a time axis; determining a plurality of first signal values of the plurality of segmented signals respectively at corresponding predetermined time points of the corresponding predetermined time periods, where the plurality of first signal values are used to represent signal values of the corresponding segmented signals at a plurality of predetermined time points of the corresponding predetermined time periods; determining waveform change frequency values corresponding to the plurality of segmented signals based on the plurality of first signal values; and determining the signal to be detected from the plurality of segmented signals based on the plurality of waveform change frequency values.

[0007] Optionally, determining the target signal feature corresponding to the signal to be detected includes: determining the signal center point corresponding to the signal to be detected, where the signal center point is used to represent the concentrated distribution point of the signal to be detected in the target domain space, and the target domain space includes at least one of the following: time domain, frequency domain, and the concentrated distribution point is the point where the distribution density of the signal to be detected is greater than a predetermined density threshold; determining the equivalent parameter corresponding to the signal to be detected according to the signal center point, where the equivalent parameter is used to represent the distribution range of the signal to be detected in the target domain space; and determining the target signal feature corresponding to the signal to be detected according to the equivalent parameter.

[0008] Optionally, obtaining multiple feature distances according to the target signal feature and the multiple cluster centers includes: obtaining an updated space according to the target signal feature and the cluster space; determining multiple updated cluster centers corresponding to the updated space; determining multiple center distances according to the multiple updated cluster centers and the multiple cluster centers, where the multiple center distances correspond to the multiple updated cluster centers one by one, and the center distance is used to represent the distance between the updated cluster center and the corresponding cluster center; and when the multiple center distances are all less than a distance threshold, determining the feature distances corresponding to the multiple updated cluster centers respectively for the target signal feature to obtain the multiple feature distances.

[0009] Optionally, determining the signal center point corresponding to the signal to be detected includes: when the signal to be detected is a continuous signal, determining the density function of the signal to be detected, where the density function is used to represent the probability value of the existence of the signal to be detected in the target interval, and the target interval includes at least one of the following: target time period, target frequency band; determining the signal cumulative value of the signal to be detected in the target interval according to the density function, where the signal cumulative value is used to represent the sum of the signal values of the signal to be detected in the target interval; and determining the signal center point corresponding to the signal to be detected according to the signal cumulative value.

[0010] Optionally, determining the signal center point corresponding to the signal to be detected includes: when the signal to be detected is a discrete signal, determining the second signal values corresponding to the signal to be detected at multiple target domain points, where the multiple target domain points include at least one of the following: multiple target time points, multiple target frequency points; and determining the signal center point corresponding to the signal to be detected according to the second signal values corresponding to the multiple target domain points.

[0011] Optionally, before retrieving the clustering space corresponding to the signal parameters of the signal to be detected, the method further includes: determining sample signals corresponding to the multiple candidate signal types respectively; determining sample features corresponding to the multiple sample signals respectively; and clustering the multiple sample features to obtain the clustering space.

[0012] According to one aspect of an embodiment of the present invention, there is provided a device for determining a signal type, including: an acquisition module configured to acquire a signal to be detected, wherein a waveform change frequency value corresponding to the signal to be detected is greater than a waveform change frequency threshold; a first determination module configured to determine a target signal feature corresponding to the signal to be detected; a retrieval module configured to retrieve a clustering space corresponding to the signal parameters of the signal to be detected, wherein the clustering space includes multiple clustering centers, and the multiple clustering centers correspond to multiple candidate signal types one by one, and the multiple candidate signal types are signal types for which a probability of belonging signal types corresponding to the signal parameters is greater than a predetermined threshold; a second determination module configured to obtain multiple feature distances based on the target signal feature and the multiple clustering centers; and a third determination module configured to determine a target signal type corresponding to the signal to be detected from the multiple candidate signal types based on the multiple feature distances.

[0013] According to one aspect of an embodiment of the present invention, there is provided an electronic device, including: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute the instructions to implement the signal type determination method according to any one of the above.

[0014] According to one aspect of an embodiment of the present invention, there is provided a computer-readable storage medium, when instructions in the computer-readable storage medium are executed by a processor of an electronic device, enabling the electronic device to execute the signal type determination method according to any one of the above.

[0015] In an embodiment of the present invention, a signal to be detected is acquired, wherein a waveform change frequency value corresponding to the signal to be detected is greater than a waveform change frequency threshold; a target signal feature corresponding to the signal to be detected is determined; a clustering space corresponding to the signal parameters of the signal to be detected is retrieved, wherein the clustering space includes a plurality of clustering centers, and the plurality of clustering centers correspond one-to-one to a plurality of candidate signal types, and the plurality of candidate signal types are signal types for which the probability of belonging signal types corresponding to the signal parameters is greater than a predetermined threshold; according to the target signal feature and the plurality of clustering centers, a plurality of feature distances are obtained; according to the plurality of feature distances, a target signal type corresponding to the signal to be detected is determined from the plurality of candidate signal types. By determining the plurality of feature distances corresponding to the target signal feature, the purpose of determining the signal type of the signal to be detected is achieved. Since the plurality of feature distances can reflect the similarity degrees of the signal to be detected with the signals of the plurality of signal types respectively, the signal type of the signal to be detected can be determined according to the feature distances, thereby solving the technical problem in the related art that when there are signals of multiple discharge types or interference signals in the detection environment, the accuracy of the determined signal type is low. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the present invention, and constitute a part of this application. The illustrative embodiments of the present invention and their descriptions are used to explain the present invention, and do not constitute an improper limitation to the present invention. In the drawings:

[0017] Figure 1 is a flowchart of a method for determining a signal type according to an embodiment of the present invention;

[0018] Figure 2 is a pulse waveform diagram provided by an alternative embodiment of the present invention;

[0019] Figure 3 is a schematic diagram of signal features of a pulse signal provided by an alternative embodiment of the present invention;

[0020] Figure 4 is a schematic diagram of equivalent parameters provided by an alternative embodiment of the present invention;

[0021] Figure 5 is a structural block diagram of a device for determining a signal type according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0023] It should be noted that the terms "first", "second", etc. in the specification, claims and above-mentioned drawings of the present invention are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present invention described here can be implemented in an order other than those illustrated or described here. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0024] Embodiment 1

[0025] According to an embodiment of the present invention, an embodiment of a method for determining a signal type 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 the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0026] Figure 1 is a flowchart of the method for determining the signal type according to an embodiment of the present invention. As Figure 1 shown, the method includes the following steps:

[0027] Step S102, obtain a signal to be detected, where the waveform change frequency value corresponding to the signal to be detected is greater than the waveform change frequency threshold.

[0028] In step S102 provided in the present application, a signal to be detected is obtained.

[0029] Among them, the signal to be detected is involved. The signal to be detected refers to the signal whose signal type needs to be identified. For example, in cable partial discharge monitoring, the signal to be detected is the partial discharge signal collected from the cable.

[0030] Among them, the waveform change frequency value is involved. The waveform change frequency value refers to the short-time average zero-crossing rate of the signal. The zero-crossing rate represents the situation where the signal value crosses the horizontal axis. When determining the partial discharge signal, the start and end times of the signal are determined according to the short-time zero-crossing law.

[0031] Among them, the waveform change frequency threshold is involved. The waveform change frequency threshold refers to a preset value used to distinguish signals of different natures. For example, in cable partial discharge detection, the waveform change frequency threshold is used to identify partial discharge signals. When the waveform change frequency of a signal exceeds the waveform change frequency threshold, the signal is determined to be a partial discharge signal.

[0032] In this step, when starting signal analysis and determining the signal type, first select the signals whose waveform change frequency values exceed the preset threshold as the signals to be detected. In the scenario of cable partial discharge monitoring, signals with high waveform change frequency values are more likely to be associated with partial discharge signals. By setting and applying the waveform change frequency threshold, low-frequency noise signals can be effectively filtered out, reducing the complexity and workload of subsequent signal analysis and processing, and improving the detection rate and recognition accuracy of specific signals such as partial discharge.

[0033] Step S104: Determine the target signal characteristics corresponding to the signal to be detected.

[0034] In step S104 provided in this application, the target signal characteristics corresponding to the signal to be detected are determined.

[0035] Among them, the target signal characteristics are involved. The target signal characteristics refer to signal attributes or parameters. For example, time-domain parameters of the signal (such as pulse width, pulse duration), frequency-domain parameters (such as main frequency, spectral bandwidth), statistical characteristics (such as mean, variance), and other characteristics that help distinguish partial discharge from other types of signals.

[0036] In this step, identify and extract the target signal characteristics of the signal to be detected. For example, time-domain characteristics, frequency-domain characteristics, and amplitude characteristics of the signal, etc. The clarification of the target signal characteristics helps optimize the signal processing algorithm, reduce unnecessary calculations, improve the subsequent analysis speed, and at the same time, can also reduce the interference of noise or non-related signals, improving the reliability of partial discharge signal detection. This is a basic step for determining the signal type.

[0037] Step S106: Retrieve the clustering space corresponding to the signal parameters of the signal to be detected. Among them, the clustering space includes multiple clustering centers, and the multiple clustering centers correspond one-to-one to multiple candidate signal types. The multiple candidate signal types are signal types whose probabilities of belonging to the signal types corresponding to the signal parameters are greater than a predetermined threshold.

[0038] In step S106 provided in this application, according to the matching result, the information type corresponding to the target information is determined.

[0039] Among them, the signal parameters are involved. The signal parameters refer to quantitative indicators that describe signal attributes or characteristics, such as frequency, amplitude, energy, zero-crossing rate, etc.

[0040] Among them, the clustering space is involved. The clustering space refers to a three-dimensional space generated by a clustering algorithm and composed of multiple clustering centers. Each clustering center corresponds to the average features of a group of similar signals and is used to classify unknown signals.

[0041] Among them, the clustering center is involved. The clustering center refers to a point or multi-dimensional vector representing the characteristics of a clustering group in the clustering space. Each clustering center corresponds to a specific signal type.

[0042] Among them, the candidate signal type is involved. The candidate signal type refers to the signal type determined according to signal parameters and having a relatively high correlation with the signal to be detected.

[0043] Among them, the attributed signal type is involved. The attributed signal type refers to the signal type of the signal to be detected.

[0044] Among them, the predetermined threshold is involved. The predetermined threshold refers to the boundary value used to distinguish signal types. If the probability that the signal type is the attributed signal type exceeds the predetermined threshold, then this signal type will be regarded as the candidate signal type.

[0045] In this step, first, determine the signal parameters of the signal to be detected. Then, retrieve the clustering space matching the signal parameters. This clustering space contains multiple clustering centers, and the multiple clustering centers correspond one-to-one with multiple candidate signal types. The multiple candidate signal types are the signal types selected from all signal types whose probability of being the attributed signal type of the signal to be detected is higher than the predetermined threshold.

[0046] Through this step, retrieving the clustering space matching the signal parameters of the signal to be detected effectively filters out low-probability signal types, reduces the number of signal types for subsequent complex analysis, thereby accelerating the signal recognition speed, and at the same time reducing the noise and uncertainty in the signal classification process and improving the classification accuracy.

[0047] Step S108: Obtain multiple feature distances based on the target signal features and multiple clustering centers.

[0048] In step S108 provided in this application, multiple feature distances are obtained.

[0049] Among them, the feature distance is involved. The feature distance refers to the distance from the target signal features to the clustering centers respectively corresponding to multiple candidate signal types.

[0050] Through this step, the target signal features of the signal to be detected are compared with the clustering center features corresponding to multiple candidate signal types respectively, and the feature distances corresponding to multiple candidate signal types are determined. These distances reflect the differences or similarities between the signal to be detected and the signal types represented by each clustering center, can distinguish partial discharge signals from interference pulses, help to accurately identify partial discharge events in a complex electromagnetic environment, and are the basis for subsequent signal identification and classification.

[0051] Step S110: Determine the target signal type corresponding to the signal to be detected from multiple candidate signal types according to multiple feature distances.

[0052] In step S110 provided in this application, the target signal type corresponding to the signal to be detected is determined from multiple candidate signal types.

[0053] Among them, the target signal type is involved. The target signal type refers to the signal type finally determined from multiple candidate signal types and can best describe the nature of the signal to be detected.

[0054] Through this step, by comparing the distances between the features of the signal to be detected and different clustering centers, the signal type of the signal to be detected is determined. By determining the target signal type in this way, it can be ensured that the signal classification decision is based on the closest match between the signal features and the historical data, thereby improving the accuracy and reliability of identification. At the same time, the signal classification can be completed in a short time, providing an immediate detection result.

[0055] Through the above steps S102 - S110, the signal to be detected is obtained, where the waveform change frequency value corresponding to the signal to be detected is greater than the waveform change frequency threshold; the target signal features corresponding to the signal to be detected are determined; the clustering space corresponding to the signal parameters of the signal to be detected is retrieved, where the clustering space includes multiple clustering centers, and multiple clustering centers correspond to multiple candidate signal types one by one, and multiple candidate signal types are signal types with the probability of belonging signal types corresponding to the signal parameters greater than a predetermined threshold; according to the target signal features and multiple clustering centers, multiple feature distances are obtained; according to multiple feature distances, the method of determining the target signal type corresponding to the signal to be detected from multiple candidate signal types, by determining multiple feature distances corresponding to the target signal features, achieves the purpose of determining the signal type of the signal to be detected. Since multiple feature distances can reflect the similarity degrees between the signal to be detected and the signals of multiple signal types respectively, the signal type of the signal to be detected can be determined according to the feature distances, thus solving the technical problem in the related art that when there are signals of multiple discharge types or interference signals in the detection environment, the accuracy of the determined signal type is low.

[0056] As an alternative embodiment, before obtaining the signal to be detected, it further includes: obtaining the original signal; determining segmented signals corresponding to multiple predetermined time periods respectively based on the original signal, where the multiple predetermined time periods are continuous on the time axis; determining multiple first signal values of the multiple segmented signals on their corresponding predetermined time periods respectively, where the multiple first signal values are used to represent the signal values corresponding to multiple predetermined time points of the corresponding segmented signals on the corresponding predetermined time periods; determining waveform change frequency values corresponding to the multiple segmented signals respectively based on the multiple first signal values; and determining the signal to be detected from the multiple segmented signals based on the multiple waveform change frequency values.

[0057] In this embodiment, the specific steps for determining the signal to be detected from the original signal are described.

[0058] Among them, the original signal is involved. The original signal refers to the signal data directly collected from the sensor or signal source without any processing. In the cable partial discharge monitoring, the original signal refers to the electrical signals or electromagnetic signals collected from the cable monitoring system or sensor, and these signals contain information about partial discharge activities, interference, and other environmental noises.

[0059] Among them, the predetermined time period is involved. The predetermined time period refers to a series of continuous time windows artificially set during signal processing for dividing the original signal into smaller and manageable segments. The length and position of these windows can be adjusted according to signal characteristics and analysis requirements.

[0060] Among them, the segmented signal is involved. The segmented signal refers to the window signals obtained by dividing the original signal according to multiple predetermined time periods and corresponding to the multiple predetermined time periods respectively.

[0061] Among them, the first signal value is involved. The first signal value refers to the signal value of the segmented signal at the predetermined time point.

[0062] Among them, the predetermined time point is involved. The predetermined time point refers to all time frames included in the corresponding predetermined time period.

[0063] In this step, first, the original signal is obtained, and then the original signal is segmented into signal segments within multiple continuous predetermined time periods, that is, the segmented signals. Next, multiple first signal values are determined within each predetermined time period to describe the instantaneous characteristics of the signal at these time points. Then, according to these first signal values, the waveform change frequency values of the segmented signals within the corresponding predetermined time periods are determined, and these frequency values reflect the dynamic characteristics of the signal waveform changing over time. Finally, based on these waveform change frequency values, the signal segments whose waveform change frequency values exceed the waveform change frequency threshold are selected from the segmented signals as the signal to be detected.

[0064] Through this step, the original signal is segmented into signal segments of multiple predetermined time periods, and the characteristics of the signal within each short time window can be focused on, which helps to identify sudden and short-term discharge activities in partial discharge monitoring. By determining the signals whose waveform change frequency values exceed the waveform change frequency threshold, the signals containing partial discharge characteristics can be quickly screened out, avoiding complex and time-consuming analysis of all the original data, thereby improving the efficiency of signal processing.

[0065] As an optional embodiment, determining the target signal characteristics corresponding to the signal to be detected includes: determining the signal center point corresponding to the signal to be detected, where the signal center point is used to represent the concentrated distribution point of the signal to be detected in the target domain space, and the target domain space includes at least one of the following: time domain, frequency domain, and the concentrated distribution point is the point where the distribution density of the signal to be detected is greater than a predetermined density threshold; determining the equivalent parameter corresponding to the signal to be detected according to the signal center point, where the equivalent parameter is used to represent the distribution range of the signal to be detected in the target domain space; and determining the target signal characteristics corresponding to the signal to be detected according to the equivalent parameter.

[0066] In this embodiment, the specific steps of determining the target signal characteristics corresponding to the signal to be detected are described.

[0067] Among them, the signal center point is involved. The signal center point refers to the concentrated distribution point of the signal in the target domain space, which can be understood as the center of gravity or average position of the signal in this space. If the signal has a concentrated distribution in the time domain or frequency domain, then the signal center point can represent the main activity area or frequency center of gravity of the signal.

[0068] Among them, the target domain space is involved. The target domain space refers to the specific dimension for signal characteristic representation or analysis, for example, the time domain, frequency domain, or any other space that can effectively represent signal characteristics.

[0069] Among them, the predetermined density threshold is involved. The predetermined density threshold refers to a preset value used to distinguish the concentrated distribution and sparse distribution of the signal. When the distribution density of the signal at a certain point exceeds this threshold, this point can be regarded as the signal center point or the concentrated area of the signal.

[0070] Among them, the equivalent parameter is involved. The equivalent parameter refers to the quantitative index that can describe the distribution range or signal characteristics of the signal in the target domain space. In the time domain or frequency domain, the equivalent parameter can be the equivalent duration, equivalent frequency bandwidth, etc. of the signal, which is used to characterize the shape or change trend of the signal.

[0071] In this step, first, determine the signal center point of the signal to be detected in the target domain space. The signal center point reflects the concentrated position of the signal in a specific domain space, usually the point with the highest signal distribution density. Then, determine the equivalent parameters of the signal to be detected based on the signal center point. These parameters quantify the distribution range or shape of the signal in the target domain space, such as the equivalent duration in the time domain, the equivalent frequency bandwidth in the frequency domain, etc. Finally, determine the target signal features corresponding to the signal to be detected based on the equivalent parameters, that is, the characteristic description of the signal. These features will be used for subsequent signal classification or recognition.

[0072] Through this step, the signal center point and equivalent parameters of the signal to be detected are determined. By analyzing the signal center point and equivalent parameter information corresponding to the signal to be detected in the target domain space, the characteristics of the signal can be described more accurately, improving the accuracy of recognition.

[0073] As an alternative embodiment, based on the target signal features and multiple cluster centers, obtain multiple feature distances, including: obtaining an updated space based on the target signal features and the cluster space; determining multiple updated cluster centers corresponding to the updated space; determining multiple center distances based on the multiple updated cluster centers and the multiple cluster centers, where the multiple center distances correspond one-to-one to the multiple updated cluster centers, and the center distance is used to represent the distance between the updated cluster center and the corresponding cluster center; when all the multiple center distances are less than the distance threshold, determine the feature distances corresponding to the multiple updated cluster centers from the target signal features, obtaining multiple feature distances.

[0074] In this embodiment, the specific steps of obtaining multiple feature distances based on the target signal features and multiple cluster centers are described.

[0075] Among them, an updated space is involved. The updated space refers to the new feature space obtained by adjusting the original cluster space according to the target signal features of the signal to be detected during the signal processing process.

[0076] Among them, an updated cluster center is involved. The updated cluster center refers to the cluster center in the updated space, which reflects the concentration trend of the data points in the adjusted feature space.

[0077] Among them, a center distance is involved. The center distance refers to the distance between the updated cluster center and the original cluster center, and this distance can be used to measure the degree of adjustment of the cluster center.

[0078] Among them, a distance threshold is involved. The distance threshold refers to a boundary value set in signal recognition and classification, which is used to determine whether the distance between the updated cluster center and the original cluster center is within a reasonable range. If the central distances between all updated cluster centers and the original cluster center are less than the distance threshold, it indicates that the updated cluster center still remains consistent or similar to the original cluster center, and the updated cluster center can be used for subsequent signal classification.

[0079] In this step, first, the clustering space is adjusted based on the target signal features of the signal to be detected to form an updated space. Then, in this updated space, new updated cluster centers are determined. Next, the central distance between the updated cluster center and the original cluster center is determined. If all central distances are less than the preset distance threshold, then the feature distance from the target signal features to the updated cluster center is further calculated. Through this step, the cluster center is dynamically adjusted to ensure the matching degree between the cluster center and the signal features, reduce misclassification and missed classification, and improve the accuracy of signal recognition.

[0080] As an optional embodiment, determining the signal center point corresponding to the signal to be detected includes: in the case where the signal to be detected is a continuous signal, determining the density function of the signal to be detected, where the density function is used to represent the probability value of the existence of the signal to be detected in the target interval, and the target interval includes at least one of the following: the target time period, the target frequency band; according to the density function, determining the signal cumulative value of the signal to be detected in the target interval, where the signal cumulative value is used to represent the sum of the signal values of the signal to be detected in the target interval; according to the signal cumulative value, determining the signal center point corresponding to the signal to be detected.

[0081] In this embodiment, the specific steps for determining the signal center point corresponding to the signal to be detected in the case where the signal to be detected is a continuous signal are described.

[0082] Among them, a continuous signal is involved. A continuous signal refers to a signal that has a continuous mathematical definition over the entire time range.

[0083] Among them, a density function is involved. The density function refers to a function of the probability of a signal appearing in a specific interval (time interval, frequency interval, etc.). For a continuous signal, the density function can be the signal intensity distribution in the time domain or the signal frequency distribution in the frequency domain. It is used to quantify the possibility of a signal appearing at different points or intervals.

[0084] Among them, a signal cumulative value is involved. The signal cumulative value refers to the accumulated result of the signal values of the signal in the target interval. For example, the sum of signal intensities, the accumulation of signal energies, or the total frequency of signal frequencies, etc. It reflects the overall activity degree of the signal in the target interval.

[0085] In this step, when the signal to be detected is a continuous signal, first, determine the density function of the signal to be detected within the target interval (such as the target time period, target frequency band). Then, determine the signal cumulative value of the signal to be detected within the target interval according to the density function. The signal cumulative value usually reflects the overall intensity or activity level of the signal within this interval. Finally, determine the signal center point of the signal to be detected according to the signal cumulative value. The signal center point describes the concentrated position of the signal within the target interval, that is, the main activity area of the signal.

[0086] As an alternative embodiment, determining the signal center point corresponding to the signal to be detected includes: in the case where the signal to be detected is a discrete signal, determining the second signal values respectively corresponding to the signal to be detected at multiple target domain points, where the multiple target domain points include at least one of the following: multiple target time points, multiple target frequency points; determining the signal center point corresponding to the signal to be detected according to the second signal values respectively corresponding to the multiple target domain points.

[0087] In this embodiment, the specific steps for determining the signal center point corresponding to the signal to be detected in the case where the signal to be detected is a discrete signal are described.

[0088] Among them, discrete signals are involved. Discrete signals refer to signals that are discrete in the time variable and have signal values only at specific time points.

[0089] Among them, the second signal value is involved. The second signal value refers to the signal value corresponding to the discrete signal at the target domain point. For time-domain analysis, the second signal value represents the signal intensity; for frequency-domain analysis, it represents the amplitude or energy of a specific frequency.

[0090] Among them, the target domain point is involved. The target domain point refers to a specific point for evaluating signal characteristics during the signal analysis process. In the time domain, the target domain point represents a time point, and in the frequency domain, the target domain point represents a frequency value.

[0091] In this step, when the signal to be detected is a discrete signal, first, determine the second signal values of the signal to be detected at multiple target domain points. For example, the values respectively corresponding to the signal at multiple time points, or the values of the signal at multiple frequency points. Then, determine the signal center point according to the multiple second signal values, that is, the concentrated position of the signal feature distribution in the target time domain or frequency domain.

[0092] As an alternative embodiment, before retrieving the clustering space corresponding to the signal parameters of the signal to be detected, it further includes: determining the sample signals respectively corresponding to multiple candidate signal types; determining the sample features respectively corresponding to the multiple sample signals; clustering the multiple sample features to obtain the clustering space.

[0093] In this embodiment, the specific steps for determining the clustering space are described.

[0094] Among them, a sample signal is involved. The sample signal refers to a representative signal selected from each candidate signal type and is used to construct a training set for subsequent signal feature extraction and clustering analysis.

[0095] Among them, sample features are involved. The sample features refer to the key attributes extracted from the sample signals and used to describe and distinguish different signal types. For example, combinations of signal parameters or high-dimensional feature vectors obtained by complex feature extraction algorithms.

[0096] In this step, first, sample signals corresponding to multiple candidate signal types are selected. The sample signals can cover the main features of the corresponding candidate signal types. Then, the corresponding sample features are extracted from the multiple sample signals respectively. Next, clustering analysis is performed on the multiple sample features to construct a clustering space. In this space, the sample signals are automatically grouped into different clusters, and each cluster represents a signal type or pattern. The cluster center reflects the concentrated features of this type of signal. Through this step, a clustering space is constructed, and sample features are extracted and analyzed separately from multiple candidate signal types in advance, which helps to more accurately distinguish different types of signals during subsequent signal recognition.

[0097] Based on the above embodiments and alternative embodiments, an alternative implementation manner is provided, which is specifically described below.

[0098] As an important device for power transmission, high-voltage cables are an important part of the power transmission and distribution system. Partial discharge (hereinafter referred to as PD) refers to an electrical discharge where the insulation between conductors is only partially bridged. There are various types of PD, such as corona discharge, internal discharge, surface discharge, etc. Different types of PD have different degrees of harm to the equipment and require different diagnostic and maintenance strategies. PD monitoring can collect PD signals in real time and analyze and process them, which is beneficial to timely discovering potential cable faults and ensuring the safe and reliable operation of the cable.

[0099] Early faults trigger weak partial discharges at the order of picocoulombs (pC). However, the electromagnetic environment at the test site is very complex and often generates noise of a comparable order of magnitude. The sources of interference include both internal interference generated by the power system itself and external electromagnetic interference from non-detection systems. The noise generated by various high-voltage electrical equipment and detection system equipment and the electromagnetic radiation generated by various carrier communications, radio communications, etc. are not independent but act together and are intertwined. That is, random white noise, pulse-type interference, periodic narrowband interference, etc. often occur simultaneously and are superimposed on the partial discharge signals, thus forming a complex and variable oscillatory wave acquisition signal. Noise can be classified into narrowband interference and broadband interference according to the frequency band, and can be classified into three categories: continuous periodic interference, pulse-type interference, and white noise interference according to the time-domain waveform characteristics.

[0100] Partial discharge pattern recognition is one of the important links in partial discharge condition monitoring. The difficulty of partial discharge recognition is relatively high, and the difficulties are as follows:

[0101] 1) In partial discharge monitoring, on the one hand, the energized operation of electrical equipment generates electromagnetic interference, and on the other hand, the partial discharge signals generated by insulation defects are usually very weak. Therefore, partial discharge signals are easily submerged in severe background noise and are not easy to detect;

[0102] 2) There are various types of partial discharges, and some discharge types have high similarity, which further increases the difficulty of partial discharge pattern recognition.

[0103] In related technologies, most of the methods for recognizing partial discharge signal patterns of distribution network cables focus on the type recognition of a single discharge source, such as the partial discharge phase spectrum diagram (Phase resolved partial discharge, PRPD) method based on statistical principles, and the improved phase spectrum recognition methods based on convolutional neural networks (BP neural networks) and support vector machines. Such methods can only accurately recognize the type of partial discharge when a single discharge source exists. When there are multiple similar partial discharges in the cable or random interference exists in the detection environment, the obtained PRPD patterns are superimposed together, and there will be a mismatch when comparing with the single discharge type in the feature database, resulting in a significant reduction in accuracy. The following specifically introduces several partial discharge signal detection methods.

[0104] 1) A method of fitting the pulse amplitude of partial discharge with a multi-parameter Weibull distribution to identify the type of multiple discharge sources. When the discharge signal comes from multiple discharge sources, the local discharge pulse amplitude distribution will have multiple inflection points, thereby judging the type of mixed discharge source. However, this method cannot identify interference. When the interference signal is large, the discharge signal is easily submerged and cannot be accurately judged.

[0105] 2) A method combining the equivalent time-frequency method and fuzzy clustering analysis. First, use the equivalent time-frequency transformation to convert the original pulse signal into the clustering mode of time-frequency (Time-Frequency, T-F) analysis, and then use fuzzy clustering to classify the data points of the T-F map, and compare with the pre-established single partial discharge defect fingerprint library to complete the classification and recognition of multi-discharge source signals. However, when the degree of the same defect is different, the distribution in the T-F mode is inconsistent, resulting in very high requirements for the fingerprint library and making the effect not ideal in the actual on-site application process.

[0106] 3) A multi-defect partial discharge signal separation based on an independent component analysis algorithm (FastICA algorithm) converts the identification of partial discharge signals from multiple sources into a single signal identification problem. However, this method also has high requirements on the noise in the test system. When the interference noise in the test environment is too large, the accuracy of the identification result will be significantly reduced.

[0107] In summary, the current partial discharge signal detection methods for distribution network cables mainly have the following problems: most methods can only identify a single partial discharge type, and cannot accurately diagnose partial discharge signals when multiple partial discharge types exist. The conventional mixed partial discharge type identification method cannot eliminate the influence of noise interference signals. The generation of these interference signals brings great difficulties to the signal identification detected by the pulse current method, which significantly increases the misjudgment rate of partial discharge signals. In addition, the required fault database is huge and the database maintenance is too frequent. Therefore, it is particularly important to identify and eliminate interference signals on site.

[0108] In view of this, a method for constructing characteristics of cable partial discharge and identifying interference pulses is provided in an optional embodiment of the present invention. By combining the analysis of characteristics of multiple types of signals, a method for judging the start and end points of pulses based on short-time energy and short-time average zero-crossing rate is used, and identification is performed with the aid of a clustering algorithm. This can improve the accuracy of partial discharge signal identification, effectively distinguish interference pulses, and solve the current problems existing in partial discharge signal detection in distribution network cables.

[0109] The following describes in detail the method steps provided by the optional implementation mode of the present invention.

[0110] S1. Obtain the signal to be detected.

[0111] A1. Get the original signal.

[0112] A2. Determine, based on the original signal, segmentation signals corresponding to a plurality of predetermined time periods.

[0113] A3. Determine a plurality of first signal values ​​of a plurality of segmentation signals in corresponding predetermined time periods.

[0114] The analysis of short-time energy is based on the change of high-frequency pulse signal amplitude over time, where the short-time energy function is:

[0115] x n (m) = x(m)w(nm)

[0116] Among them, x n (m) is the amplitude of the windowed signal to be detected (the same as the segmented signal mentioned above) at point m (the same as the first signal value mentioned above), x(m) is the amplitude of the original signal at point m, w(nm) is the window function, where n is the window position and m is the time index.

[0117] The square of the window function h(n) is:

[0118] h(n) = w 2 (n)

[0119] The short - time energy E n is:

[0120]

[0121] A4. Determine the waveform change frequency values corresponding to the multiple segmented signals based on multiple first signal values.

[0122] Judge the zero - crossing rate of the calculation signal based on the short - time energy function. The zero - crossing rate represents the situation where the signal crosses the horizontal axis. For a continuous signal, the zero - crossing rate is the situation where the time - domain waveform passes through the horizontal axis; for a discrete signal, that is, adjacent sampled values have different algebraic signs, which is the number of times the sample points change signs.

[0123] Considering that the pulse signal is a non - stationary signal, when using the zero - crossing rate as a certain judgment index, only its short - time average zero - crossing rate (the same as the above waveform change frequency) can be considered. The short - time average zero - crossing rate Z n is:

[0124]

[0125] where N is the number of multiple first signal values, and the sign function sgn[x(n)] is:

[0126]

[0127] Determine the short - time average zero - crossing rate Z n ′ of the introduced window function is:

[0128]

[0129] Furthermore, determine the short - time average zero - crossing rate Z n ″ under the sliding window of length N is:

[0130]

[0131] As shown in the above formula, when the signs of two adjacent sampled values are the same, no zero - crossing occurs, and when the signs of two adjacent sampled values are opposite, sgn[x(n)] - sgn[x(-1)] = 2. Therefore, dividing the sum by 2N can obtain the average zero - crossing rate.

[0132] A5. Determine the signal to be detected from the multiple segmented signals based on multiple waveform change frequency values.

[0133] For a non - stationary signal, calculate its short - time average zero - crossing rate, and then judge the starting and ending points of the pulse.

[0134] S2. Determine the target signal characteristics corresponding to the signal to be detected.

[0135] Pulses excited by different discharge sources have different waveform characteristics. Figure 2 It is the pulse waveform diagram provided by an alternative embodiment of the present invention. As Figure 2 shown, where A and B are pulses excited by two discharge sources, and their waveforms are significantly different, but there are also certain commonalities. In most cases, partial discharges excited by cable defects are also significantly different from the waveforms of noise signals. The partial discharge signal is different from the non-partial discharge signal in terms of pulse waveform. Utilizing this point, the separation of discharge pulses and interference pulses can be achieved, thereby realizing the extraction of partial discharge pulses from the acquired signals. Study the conversion mechanism between the primary side voltage and current and the partial discharge signal, study the weak signal extraction algorithm and the mechanism of high-precision sensing technology, analyze the propagation mechanism of interference pulses in distribution cables, and analyze the signal characteristics reaching typical positions. Figure 3 It is the schematic diagram of the signal characteristics of the pulse signal provided by an alternative embodiment of the present invention. As Figure 3 shown, P is the pulse peak value of the signal, Tr is the rise time of the pulse, Tf is the fall time of the pulse, Tw is the pulse width, and Tl is the pulse duration.

[0136] B1. Construct characteristic indexes of time-domain parameters such as pulse rise time, pulse fall time, pulse width, and pulse duration.

[0137] (1) The pulse rise time Tr, that is, the pulse rising edge, refers to the time interval between the main pulse of the partial discharge waveform rising from 10% of the pulse peak value to 90% of the peak value.

[0138] (2) The pulse fall time Tf, that is, the pulse falling edge, is opposite to the pulse rise time, and refers to the time interval between the main pulse of the partial discharge waveform starting to fall from 90% of the pulse peak value to 10% of the peak value.

[0139] (3) The pulse width Tw refers to the time interval between two points at 50% of the peak value of the partial discharge main pulse.

[0140] (4) The pulse duration Tl refers to the time interval from the occurrence of the partial discharge to the end of the oscillation. The pulse duration characterizes the effective duration of the discharge.

[0141] B2. Construct characteristic indexes of amplitude-domain parameters such as pulse peak value index, pulse mean value, pulse absolute mean discharge effective value, pulse crest factor, and pulse waveform factor.

[0142] (1) Pulse peak value P: The pulse peak value refers to the maximum amplitude of the main pulse in the partial discharge waveform. The polarity of the peak value is related to the detection circuit, and its magnitude is related to the DC voltage application magnitude and the oscillating wave frequency. The oscillating wave high-speed data acquisition device collects the voltage signal of the sampling resistor, but essentially collects the current signal. Its absolute value represents the maximum discharge current of the partial discharge and also represents the maximum discharge amount of the partial discharge.

[0143] (2) Pulse mean value μ m , representing the average current of the partial discharge, and its expression is:

[0144]

[0145] Among them, z(i) represents the current value measured at a certain specific moment or position during the partial discharge process, and q represents the total number of data points participating in the calculation of the average value.

[0146] (3) Absolute pulse mean value μ a , which refers to the mean value of the absolute value of the discharge current, and its expression is:

[0147]

[0148] (4) Discharge effective value D, which refers to the root mean square value of the discharge current, and its expression is:

[0149]

[0150] The product of the average discharge current and the detection resistor represents the average discharge power. Since the resistance of the detection circuit is fixed, therefore, the resistance value can be omitted, that is, the average discharge power can be expressed by the effective value.

[0151] (5) Pulse peak factor CF (Crest Factor), which refers to the ratio of the main pulse peak value of the partial discharge to the partial discharge effective value, and its expression is:

[0152]

[0153] Among them, P represents the main pulse peak value of the partial discharge. The peak factor is a relative ratio and is not affected by the absolute discharge amount, reflecting the relative size of the spike.

[0154] (6) Pulse waveform factor FF (Form Factor), which refers to the ratio of the partial discharge effective value to the partial discharge average value, and its expression is:

[0155]

[0156] B3. Study the T-F parameter characteristics, energy or amplitude distribution characteristics, pulse first wave polarity characteristics, etc. of the partial discharge of the distribution cable and on-site interference.

[0157] The commonly used and traditional spectral transformation tool, the Fourier transform, has played an important role in the field of stationary signal processing. For stationary signals, the characteristic spectral peaks and characteristic frequencies can be obtained through the Fourier transform. However, the Fourier transform is a global transformation that requires all the time-domain information of the signal. That is to say, the spectrum is the superposition of time from negative infinity to positive infinity and cannot reflect the time-domain information of the spectral values. Therefore, it is difficult for the Fourier transform to analyze non-stationary and short-time signals and cannot reflect the local properties of time-varying short-time signals. For partial discharge signals, other time-frequency transformation methods need to be used to extract their characteristics.

[0158] In the field of signal and system research, the time center of gravity and width of a signal in the time domain, and the frequency center and width in the frequency domain are important characteristic quantities. The time center of gravity and width describe the central position of the signal in the time domain and its extension in the time domain, while the frequency center and frequency width describe the central position of the signal in the frequency domain and its extension in the frequency domain.

[0159] Designing equivalent time-frequency calculation metrics mainly includes two parts: designing equivalent time-frequency calculation metrics for continuous signals and equivalent time-frequency calculation metrics for discrete signals.

[0160] For continuous signals (the same as the above continuous signals), design equivalent time width and frequency width calculation methods as effective characterizations of partial discharge pulses.

[0161] Under the excitation of oscillating waves with different frequencies and amplitudes, the amplitudes of partial discharges can vary greatly, ranging from dozens of pC, hundreds of pC to thousands of pC or even over ten thousand pC. According to the statistics of the knowledge base, most of the partial discharge peaks are below 5000 pC.

[0162] The following details the definition of equivalent time-frequency (the same as the above equivalent parameters).

[0163] Let the time pulse waveform be s(t), where t is the sampling time and t ∈ [0, T], and T is the time elapsed from the first appearance of the signal to the end of the signal.

[0164] First, normalize (standardize) the signal s(t) to obtain the normalized signal function S N (t), and its expression is:

[0165]

[0166] The normalization function is regarded as the density function of the signal s(t) in the time domain. Using the density function, the time expectation of s(t) can be calculated by the method of probability central moments, that is, the time "center of gravity" (t0, the same as the above signal center point) of the normalized signal S N (t) is obtained.

[0167] Figure 4 It is a schematic diagram of equivalent parameters provided by an alternative embodiment of the present invention. As Figure 4 shown, after obtaining the time centroid, the time-domain equivalent time length δ of a single waveform can be calculated by the method of standard deviation T and the frequency-domain equivalent bandwidth δ F . Among them, the time-domain equivalent time length δ of a single waveform T is:

[0168]

[0169] The frequency-domain equivalent bandwidth δ of a single waveform F is:

[0170]

[0171] Among them, f is the sampling frequency, and S N (f) is the Fourier transform of the normalized signal s(t).

[0172] The time width v T and the frequency width δ F are the concentrated condensation of the signal characteristics, respectively reflecting the degree of expansion of the signal around the time centroid t and the frequency centroid, and also reflecting the weighted average characteristics of the pulse signal in the time domain and the frequency domain.

[0173] For the oscillatory wave pulse (the same as the above discrete signal), the acquired signal is a discrete time series. According to the positioning and calculation method of equivalent time-frequency, after appropriate conversion, it is used as an effective characterization of the partial discharge pulse.

[0174] For an abnormal signal s(t i ), i = 0, 1, 2,..., n, the calculation method of its equivalent time width and frequency width is introduced in detail below.

[0175] Its time centroid T0 is:

[0176]

[0177] The frequency centroid is F0:

[0178]

[0179] Among them, f i is the sampling point (the same as the above target domain point), and S(f i ) is the Fourier transform of the abnormal signal s(t i ), i = 0, 1, 2,..., n.

[0180] The equivalent time length is δ T :

[0181]

[0182] The equivalent bandwidth is δ F :

[0183]

[0184] For each discharge pulse waveform, calculated according to the above formula, it will correspond to a vector (δ T , δ F ). Project the collected signal pulses onto the two-dimensional plane composed of equivalent time and equivalent bandwidth, that is, form the T-F mode. The calculation of equivalent time-frequency includes the entire signal, condenses the information of the discharge signal, and can effectively characterize the time-frequency characteristics of the signal.

[0185] S3. Retrieve the clustering space corresponding to the signal parameters of the signal to be detected.

[0186] Based on the above signal characteristics and index calculation results, use the FCM clustering algorithm to identify cable interference pulses. Based on the defined pulse start and end points, combined with the pulse index characteristics, equivalent pulse time-frequency characteristics, etc. constructed in the above steps, use the FCM clustering algorithm to identify partial discharge pulses and interference pulses, and obtain the clustering centers of different clusters.

[0187] S4. According to the target signal characteristics and multiple clustering centers, obtain multiple characteristic distances.

[0188] In the case of fuzzy classification, in order to obtain the optimal classification, construct an objective function similar to the sum-of-squared-errors criterion as:

[0189]

[0190] Among them, J m is the weighted sum of the squares of the distances from all samples to the clustering centers; c is the number of clustering centers; y is the number of samples; u ej is the membership degree of x to v e ; v e is the clustering center of the e-th class; m is the fuzzy index. During the clustering process, by assigning appropriate membership degrees to each sample, the objective function J m is minimized.

[0191] S5. According to multiple characteristic distances, determine the target signal type corresponding to the signal to be detected from multiple candidate signal types.

[0192] By using the FCM clustering algorithm to identify cable interference pulses, the signals are automatically classified according to the signal characteristics, reducing manual intervention, realizing intelligent processing, and improving the recognition efficiency and accuracy.

[0193] Through the above optional implementation manners, at least the following beneficial effects can be achieved:

[0194] (1) The optional embodiments of the present invention not only focus on traditional time-domain and amplitude-domain parameter characteristics, but also deeply study T-F parameter characteristics, energy or amplitude distribution characteristics, and pulse first-wave polarity characteristics, etc., forming an identification system with multi-feature fusion. This feature analysis method that synthesizes multiple dimensions can capture the subtle differences between partial discharge signals and interference pulses more accurately, thus significantly improving the accuracy of identification;

[0195] (2) Innovatively combines short-time energy and short-time average zero-crossing rate to determine the start and end points of pulses. Short-time energy can reflect the energy distribution of the signal within the time window, while the short-time average zero-crossing rate can reflect the change frequency of the signal waveform. By combining these two features, this method can more accurately locate pulse signals and reduce the situations of misjudgment and missed judgment. This optimized pulse detection method not only improves the accuracy of identification, but also provides a more reliable data basis for subsequent clustering identification;

[0196] (3) By introducing the fuzzy C-means clustering algorithm (FCM, fuzzy C-means clustering algorithm), this method realizes the intelligent identification of cable interference pulses. Compared with traditional rule-based or threshold-based identification methods, the clustering algorithm can automatically classify according to signal characteristics without the need to manually set thresholds or rules. This not only improves the automation degree of identification, but also enhances the ability of the algorithm to distinguish different types and intensities of interference pulses.

[0197] It should be noted that for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be carried out in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0198] Through the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be implemented by means of software plus a necessary general hardware platform. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, in essence, 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 disc), including several instructions for causing 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.

[0199] Embodiment 2

[0200] According to an embodiment of the present invention, there is also provided an apparatus for implementing the method for determining the signal type described above. Figure 5 It is a structural block diagram of the apparatus for determining the signal type according to an embodiment of the present invention, as Figure 5 shown. The apparatus includes: an acquisition module 502, a first determination module 504, a retrieval module 506, a second determination module 508, and a third determination module 510. The following provides a detailed description of the apparatus.

[0201] The acquisition module 502 is configured to acquire a signal to be detected, where the waveform change frequency value corresponding to the signal to be detected is greater than the waveform change frequency threshold; the first determination module 504 is connected to the above-mentioned acquisition module 502 and is configured to determine the target signal feature corresponding to the signal to be detected; the retrieval module 506 is connected to the above-mentioned first determination module 504 and is configured to retrieve a clustering space corresponding to the signal parameters of the signal to be detected, where the clustering space includes a plurality of clustering centers, and the plurality of clustering centers correspond to a plurality of candidate signal types one by one, and the plurality of candidate signal types are signal types whose probability of belonging signal types corresponding to the signal parameters is greater than a predetermined threshold; the second determination module 508 is connected to the above-mentioned retrieval module 506 and is configured to obtain a plurality of feature distances based on the target signal feature and the plurality of clustering centers; the third determination module 510 is connected to the above-mentioned second determination module 508 and is configured to determine the target signal type corresponding to the signal to be detected from the plurality of candidate signal types based on the plurality of feature distances.

[0202] It should be noted here that the above-mentioned acquisition module 502, first determination module 504, retrieval module 506, second determination module 508, and third determination module 510 correspond to steps S102 to S110 in the method for determining the signal type. The examples and application scenarios implemented by the plurality of modules and the corresponding steps are the same, but are not limited to the content disclosed in the above-mentioned Embodiment 1.

[0203] Embodiment 3

[0204] According to another aspect of an embodiment of the present invention, there is also provided an electronic device, including: a processor; a memory for storing processor-executable instructions, where the processor is configured to execute the instructions to implement the method for determining the signal type in any one of the above.

[0205] Embodiment 4

[0206] According to another aspect of an embodiment of the present invention, there is also provided a computer-readable storage medium. When the instructions in the computer-readable storage medium are executed by the processor of the electronic device, the electronic device can execute the method for determining the signal type in any one of the above.

[0207] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0208] In the above embodiments of the present invention, the descriptions of the respective embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0209] In several embodiments provided by the present application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are merely illustrative. For example, the division of the units can be a logical function division. In actual implementation, there can be other division methods. For example, 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 displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces. The indirect coupling or communication connection of units or modules can be in electrical or other forms.

[0210] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0211] In addition, in each embodiment of the present invention, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0212] If the above-mentioned integrated units are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an 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 for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The foregoing storage medium includes: USB flash drives, read-only memories (ROMs), random access memories (RAMs), mobile hard disks, magnetic disks, or optical disks and other various media that can store program codes.

[0213] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and modifications can be made, and these improvements and modifications should also be regarded as the protection scope of the present invention.

Claims

1. A method for determining a signal type, characterized in that, Including: Obtain a signal to be detected, wherein a waveform change frequency value corresponding to the signal to be detected is greater than a waveform change frequency threshold value; Determine a target signal feature corresponding to the signal to be detected; Retrieve a clustering space corresponding to signal parameters of the signal to be detected, wherein the clustering space includes a plurality of clustering centers, and the plurality of clustering centers are in one-to-one correspondence with a plurality of candidate signal types, and the plurality of candidate signal types are signal types for which a probability that they are attribution signal types corresponding to the signal parameters is greater than a predetermined threshold value; Obtain a plurality of feature distances according to the target signal feature and the plurality of clustering centers; Determine a target signal type corresponding to the signal to be detected from the plurality of candidate signal types according to the plurality of feature distances.

2. The method according to claim 1, wherein Before obtaining the signal to be detected, further including: Obtain an original signal; Determine segmented signals corresponding to a plurality of predetermined time periods according to the original signal, wherein the plurality of predetermined time periods are continuous on a time axis; Determine a plurality of first signal values of the plurality of segmented signals on corresponding predetermined time periods, wherein the plurality of first signal values are used to represent signal values corresponding to the plurality of segmented signals at a plurality of predetermined time points on the corresponding predetermined time periods respectively; Determine waveform change frequency values corresponding to the plurality of segmented signals according to the plurality of first signal values; Determine the signal to be detected from the plurality of segmented signals according to a plurality of waveform change frequency values.

3. The method according to claim 1, characterized in that, The determining the target signal feature corresponding to the signal to be detected includes: Determine a signal center point corresponding to the signal to be detected, wherein the signal center point is used to represent a concentrated distribution point of the signal to be detected in a target domain space, and the target domain space includes at least one of the following: time domain, frequency domain, and the concentrated distribution point is a point where a distribution density of the signal to be detected is greater than a predetermined density threshold value; Determine an equivalent parameter corresponding to the signal to be detected according to the signal center point, wherein the equivalent parameter is used to represent a distribution range of the signal to be detected in the target domain space; Determine the target signal feature corresponding to the signal to be detected according to the equivalent parameter.

4. The method according to claim 1, wherein The obtaining the plurality of feature distances according to the target signal feature and the plurality of clustering centers includes: Obtain an updated space according to the target signal feature and the clustering space; Determine a plurality of updated clustering centers corresponding to the updated space; Determine a plurality of center distances according to the plurality of updated clustering centers and the plurality of clustering centers, wherein the plurality of center distances are in one-to-one correspondence with the plurality of updated clustering centers, and the center distance is used to represent a distance between the updated clustering center and the corresponding clustering center; When all the plurality of center distances are less than a distance threshold value, determine feature distances corresponding to the plurality of updated clustering centers from the target signal feature to obtain the plurality of feature distances.

5. The method according to claim 3, characterized in that, The determining the signal center point corresponding to the signal to be detected includes: When the signal to be detected is a continuous signal, determine the probability density function of the signal to be detected, where the probability density function is used to represent the probability value that the signal to be detected exists in the target interval, and the target interval includes at least one of the following: a target time period, a target frequency band; According to the probability density function, determine the signal accumulation value of the signal to be detected in the target interval, where the signal accumulation value is used to represent the sum of the signal values of the signal to be detected in the target interval; According to the signal accumulation value, determine the signal center point corresponding to the signal to be detected.

6. The method according to claim 3, wherein The determination of the signal center point corresponding to the signal to be detected includes: When the signal to be detected is a discrete signal, determine the second signal values respectively corresponding to the signal to be detected at multiple target domain points, where the multiple target domain points include at least one of the following: multiple target time points, multiple target frequency points; According to the second signal values respectively corresponding to the multiple target domain points, determine the signal center point corresponding to the signal to be detected.

7. The method according to any one of claims 1 to 6, characterized in that Before retrieving the clustering space corresponding to the signal parameters of the signal to be detected, it further includes: Determine the sample signals respectively corresponding to the multiple candidate signal types; Determine the sample features respectively corresponding to the multiple sample signals; Cluster the multiple sample features to obtain the clustering space.

8. A device for determining a signal type, characterized in that, It includes: An acquisition module, configured to acquire a signal to be detected, where the waveform change frequency value corresponding to the signal to be detected is greater than a waveform change frequency threshold; A first determination module, configured to determine the target signal feature corresponding to the signal to be detected; A retrieval module, configured to retrieve the clustering space corresponding to the signal parameters of the signal to be detected, where the clustering space includes multiple cluster centers, and the multiple cluster centers correspond one-to-one to multiple candidate signal types, and the multiple candidate signal types are the signal types whose probability of belonging signal types corresponding to the signal parameters is greater than a predetermined threshold; A second determination module, configured to obtain multiple feature distances according to the target signal feature and the multiple cluster centers; A third determination module, configured to determine the target signal type corresponding to the signal to be detected from the multiple candidate signal types according to the multiple feature distances.

9. An electronic device, characterized in that, It includes: A processor; A memory for storing executable instructions of the processor; Wherein, the processor is configured to execute the instructions to implement the method for determining the signal type according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, When the instructions in the computer-readable storage medium are executed by the processor of the electronic device, the electronic device can execute the method for determining the signal type according to any one of claims 1 to 7.

Images