Feature extraction method, low-voltage electrical apparatus life prediction method, device, and medium

Abstract

The application provides a feature extraction method, a low-voltage electrical appliance life prediction method, a device and a medium, and relates to the technical field of low-voltage electrical appliances. The method comprises the following steps: acquiring working waveforms of a first electrical appliance in a plurality of opening and closing cycles in a historical time period, each working waveform comprising at least a voltage waveform and a current waveform; and extracting target feature data of each working waveform, wherein the target feature data comprises wavelet packet features and / or arc parameters. According to the embodiments of the application, the extracted target feature data can represent the relevant working parameters of the first electrical appliance, and when the target feature data is applied to the residual life prediction of the first electrical appliance, the residual life parameters of the first electrical appliance can be predicted more accurately.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of low-voltage electrical apparatus, and in particular relates to a feature extraction method, a low-voltage electrical apparatus life prediction method, a device and a medium. BACKGROUND

[0002] Low-voltage electrical apparatus generally refers to electrical apparatus working at an alternating voltage of 1200V or a direct voltage of 1500V or below, and is widely used in industrial production and life. However, in some application scenarios with high requirements for device operation stability and safety, such as elevators, high-speed trains and new energy, if the low-voltage electrical apparatus cannot be replaced in time before failure, it may cause great loss and inconvenience to enterprise production and the public, and therefore it is necessary to predict the remaining electrical life of the low-voltage electrical apparatus.

[0003] At present, the remaining service life of a low-voltage electrical apparatus is generally estimated according to the installation time of the low-voltage electrical apparatus.

[0004] It can be seen that the existing method for estimating the remaining life of a low-voltage electrical apparatus is relatively simple and has the problem of inaccurate estimation. SUMMARY

[0005] The present application aims to solve the above problems in the prior art, and provides a feature extraction method, a low-voltage electrical apparatus life prediction method, a device and a medium. When the extracted target feature data is applied to the remaining life prediction of a first electrical apparatus, the remaining life parameters of the first electrical apparatus can be predicted more accurately.

[0006] To achieve the above object, the technical solutions adopted by the embodiments of the present application are as follows:

[0007] In a first aspect, the present application provides a feature extraction method, comprising:

[0008] obtaining working waveforms of a first electrical apparatus in a plurality of opening and closing cycles in a historical time period, each of the working waveforms comprising at least: a voltage waveform and a current waveform;

[0009] extracting target feature data of each of the working waveforms, the target feature data comprising: wavelet packet features, and / or, arc parameters.

[0010] In an optional implementation, the target feature data comprises: arc parameters, the arc parameters comprising: an arc energy proportion; and the extracting of the target feature data of each of the working waveforms comprises:

[0011] calculating arc energy corresponding to each of the working waveforms according to the voltage waveform and the current waveform corresponding to each of the working waveforms;

[0012] Based on the preset phase angle interval and the arcing energy corresponding to each of the working waveforms, a preset clustering algorithm is used to cluster each of the working waveforms to obtain a clustering result.

[0013] According to the clustering result, the arcing energy proportion corresponding to the first electric appliance is calculated.

[0014] In an optional implementation, the calculation of the arcing energy proportion corresponding to the first electric appliance according to the clustering result includes:

[0015] According to the clustering result, the number of working waveforms in the plurality of working waveforms that meet a preset arcing energy threshold range is determined.

[0016] According to the number of waveforms and the total number of the plurality of working waveforms, the arcing energy proportion corresponding to the first electric appliance is calculated.

[0017] In an optional implementation, the target feature data includes wavelet packet features, and the extraction of the target feature data of each of the working waveforms includes:

[0018] Each of the working waveforms is subjected to wavelet packet decomposition to obtain a plurality of waveband waveforms corresponding to each of the working waveforms after decomposition, wherein each of the waveband waveforms corresponds to a different frequency interval.

[0019] The frequency band energy of each waveband waveform corresponding to each of the working waveforms and the Shannon entropy of each of the working waveforms are calculated.

[0020] In an optional implementation, the calculation of the Shannon entropy of each of the working waveforms includes:

[0021] According to the frequency band energy of each of the waveband waveforms corresponding to each of the working waveforms, a normalized frequency band energy coefficient of each of the waveband waveforms in each of the working waveforms is calculated.

[0022] According to the normalized frequency band energy coefficient of each of the waveband waveforms in each of the working waveforms, the Shannon entropy of each of the working waveforms is calculated.

[0023] In an optional implementation, the preset clustering algorithm includes but is not limited to one or more of a K-means clustering algorithm, a mean shift clustering algorithm, and a hierarchical clustering algorithm.

[0024] In an optional implementation, the target feature data further includes arcing time and / or arcing power, contact voltage, contact current, and contact resistance.

[0025] In a second aspect, the present application provides a low-voltage electric appliance life prediction method, including:

[0026] inputting target feature data corresponding to the to-be-predicted electrical appliance into the life prediction model to predict and obtain a remaining life parameter of the to-be-predicted electrical appliance, wherein the target feature data corresponding to the to-be-predicted electrical appliance is obtained by using the feature extraction method in any one of the preceding embodiments, and the first electrical appliance is the to-be-predicted electrical appliance.

[0027] The life prediction model is trained according to a training sample data set, the training sample data set includes target feature data corresponding to a plurality of sample electrical appliances, and a corresponding remaining life parameter is labeled, the target feature data corresponding to the plurality of sample electrical appliances is obtained by using the feature extraction method in any one of the preceding embodiments, and the first electrical appliance is the plurality of sample electrical appliances.

[0028] In a third aspect, the present application provides a feature extraction device, comprising:

[0029] The acquisition module is configured to acquire working waveforms of the first electrical appliance in a plurality of opening and closing cycles in a historical time period, and each working waveform includes at least a voltage waveform and a current waveform.

[0030] The extraction module is configured to extract target feature data of each working waveform, and the target feature data includes wavelet packet features and / or arc parameters.

[0031] In an optional embodiment, the target feature data includes arc parameters, and the arc parameters include an arc energy proportion; and the extraction module is specifically configured to calculate arc energy corresponding to each working waveform according to a voltage waveform and a current waveform corresponding to each working waveform.

[0032] Based on a preset phase angle interval and the arc energy corresponding to each working waveform, a preset clustering algorithm is used to cluster each working waveform to obtain a clustering result.

[0033] According to the clustering result, the arc energy proportion corresponding to the first electrical appliance is calculated.

[0034] In an optional embodiment, the extraction module is specifically configured to determine, according to the clustering result, a waveform number of working waveforms in the plurality of working waveforms that meet a preset arc energy threshold range.

[0035] According to the waveform number and a total number of the plurality of working waveforms, the arc energy proportion corresponding to the first electrical appliance is calculated.

[0036] In an optional embodiment, the target feature data includes wavelet packet features, and the extraction module is specifically configured to perform wavelet packet decomposition on each working waveform to obtain a plurality of waveband waveforms corresponding to each working waveform after decomposition, wherein each waveband waveform corresponds to a different frequency interval.

[0037] The frequency band energy of each of the wave band waveforms corresponding to each of the working waveforms is calculated, and the Shannon entropy of each of the working waveforms is calculated.

[0038] In an optional implementation, the extraction module is specifically configured to calculate a normalized frequency band energy coefficient of each of the wave band waveforms in each of the working waveforms according to the frequency band energy of each of the wave band waveforms corresponding to each of the working waveforms.

[0039] The Shannon entropy of each of the working waveforms is calculated according to the normalized frequency band energy coefficient of each of the wave band waveforms in each of the working waveforms.

[0040] In an optional implementation, the preset clustering algorithm includes at least one of the following: a K-means clustering algorithm, a mean shift clustering algorithm, and a hierarchical clustering algorithm.

[0041] In an optional implementation, the target feature data further includes at least one of the following: an arcing time, an arcing power, a contact voltage, a contact current, and a contact resistance.

[0042] In a fourth aspect, the present application provides a low-voltage electrical appliance life prediction device, comprising:

[0043] A prediction module is configured to input target feature data corresponding to a to-be-predicted electrical appliance into a life prediction model to predict and obtain a remaining life parameter of the to-be-predicted electrical appliance, wherein the target feature data corresponding to the to-be-predicted electrical appliance is obtained by using the steps of any one of the feature extraction methods in the foregoing embodiments, and the first electrical appliance is the to-be-predicted electrical appliance.

[0044] The life prediction model is trained according to a training sample data set, the training sample data set includes target feature data corresponding to a plurality of sample electrical appliances and labels corresponding remaining life parameters, the target feature data corresponding to the plurality of sample electrical appliances is obtained by using the steps of any one of the feature extraction methods in the foregoing embodiments, and the first electrical appliance is the plurality of sample electrical appliances.

[0045] In a fifth aspect, the present application provides an electronic device, comprising a processor, a storage medium, and a bus, the storage medium stores machine readable instructions executable by the processor, when the electronic device is running, the processor and the storage medium communicate through the bus, the processor executes the machine readable instructions to execute the steps of any one of the feature extraction methods in the foregoing embodiments or the steps of the low-voltage electrical appliance life prediction method in the foregoing embodiments.

[0046] In a sixth aspect, the present application provides a computer readable storage medium, wherein a computer program is stored on the computer readable storage medium, and when the computer program is executed by a processor, the steps of the feature extraction method according to any one of the preceding embodiments or the steps of the low-voltage electrical appliance life prediction method according to the preceding embodiments are performed.

[0047] The present application has the following beneficial effects:

[0048] In the feature extraction method, the low-voltage electrical appliance life prediction method, the device and the medium provided by the present application, the working waveforms of the first electrical appliance in a plurality of opening and closing cycles in a historical time period are obtained, and each working waveform at least includes a voltage waveform and a current waveform. Target feature data of each working waveform is extracted, and the target feature data includes wavelet packet features and / or arc parameters. By using the present application, the extracted target feature data can represent the relevant working parameters of the first electrical appliance, and when the target feature data is applied to the remaining life prediction of the first electrical appliance, the remaining life parameters of the first electrical appliance can be predicted more accurately. BRIEF DESCRIPTION OF DRAWINGS

[0049] In order to more clearly illustrate the technical solutions of the present application, the following will briefly introduce the drawings needed in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can also be obtained without creative labor.

[0050] Figure 1 A flowchart of a feature extraction method provided by the present application embodiment is shown in the figure;

[0051] Figure 2 A flowchart of another feature extraction method provided by the present application embodiment is shown in the figure;

[0052] Figure 3 A flowchart of another feature extraction method provided by the present application embodiment is shown in the figure;

[0053] Figure 4 A curve diagram of the arc energy proportion and the number of operations corresponding to a first electrical appliance provided by the present application embodiment is shown in the figure;

[0054] Figure 5 A flowchart of another feature extraction method provided by the present application embodiment is shown in the figure;

[0055] Figure 6 A wavelet packet decomposition diagram provided by the present application embodiment is shown in the figure;

[0056] Figure 7A flowchart of another feature extraction method provided by the embodiments of the present application is shown in FIG. 2;

[0057] Figure 8 A functional module diagram of a feature extraction device provided by the embodiments of the present application is shown in FIG. 3;

[0058] Figure 9 An electronic device structure provided by the embodiments of the present application is shown in FIG. 4. DETAILED DESCRIPTION

[0059] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments of the present application. The components of the embodiments of the present application described and shown in the drawings can be arranged and designed in various different configurations.

[0060] Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work are within the scope of protection of the present application.

[0061] It should be noted that: similar reference numerals and letters represent similar items in the following drawings, thus, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. Before introducing the present application, in order to better understand the present application, first, the related terms involved in the present application are explained, and the specific content is as follows.

[0062] Low-voltage electrical apparatus: generally refers to electrical apparatus working under AC voltage 1200V or DC voltage 1500V or below. Common low-voltage apparatuses include switches, fuses, contactors, leakage protectors and relays, etc. When electrical lines are installed, low-voltage apparatuses are used to connect the power supply and the load (such as a motor) through wires, so as to realize the control functions of turning on, turning off and protecting the load.

[0063] Arcing time: taking a relay as an example, it refers to the time from the opening of the moving contact to the extinguishing of the arc.

[0064] Arcing power: the product of the voltage across the low-voltage electrical apparatus and the current across the low-voltage electrical apparatus when the low-voltage electrical apparatus appears an arc during the arcing time.

[0065] Contact voltage: the voltage across the low-voltage electrical apparatus during the closing of the low-voltage electrical apparatus.

[0066] Contact current: The current across a low-voltage electrical appliance during the closing process.

[0067] Contact resistance: The ratio of the effective value of the contact voltage to the effective value of the contact current.

[0068] Arc energy: When a low-voltage electrical appliance is de-energized (e.g., when the circuit breaker is opened), an electric arc is generated between the contacts. The integral of the arc power over the arcing time is the arc energy.

[0069] Phase angle: During the transmission of alternating current, voltage and current exhibit phase changes between 0 degrees and 360 degrees.

[0070] Existing methods for estimating the remaining service life of low-voltage electrical appliances generally rely on a rough estimate based on the installation time of the appliances. Therefore, existing methods for predicting the service life of low-voltage electrical appliances suffer from inaccurate predictions.

[0071] In view of this, embodiments of this application provide a feature extraction method. By applying this feature extraction method, target feature data of the electrical appliance to be predicted can be extracted. Based on the target feature data, the remaining life parameters of the electrical appliance to be predicted can be predicted more accurately, so that maintenance personnel can replace the electrical appliance to be predicted in a timely manner according to the remaining life parameters, thereby ensuring the normal operation of the line.

[0072] Figure 1 This is a flowchart illustrating a feature extraction method provided in an embodiment of this application. The executing entity of this method can be a first electrical appliance, specifically a processing unit within the first electrical appliance that has data processing capabilities, such as a processor, etc., which is not limited here. Optionally, the executing entity of this method can also be an electronic device such as a computer or server, which may vary depending on the actual application scenario.

[0073] like Figure 1 As shown, the feature extraction method may include:

[0074] S101. Obtain the operating waveforms of the first electrical appliance during multiple opening and closing cycles within a historical time period. Each operating waveform includes a voltage waveform and a current waveform.

[0075] One opening and closing cycle refers to the working cycle of the first electrical appliance from a power-off state to a power-on state and then from a power-on state to a power-off state. Alternatively, it can refer to the working cycle of the first electrical appliance from a power-on state to a power-off state and then from a power-off state to a power-on state. There is no limitation here. The historical time period can be one week, three days, one day, etc., which is not limited here.

[0076] Optionally, the first electrical apparatus can be a low-voltage electrical apparatus, or can be another type of electrical apparatus, which is not limited herein. If it is a low-voltage electrical apparatus, the low-voltage electrical apparatus can include, but is not limited to, a fuse, a contactor, a leakage protector, a relay, and the like. For the low-voltage electrical apparatus, the working waveforms of the low-voltage electrical apparatus in the plurality of opening and closing cycles within the historical time period can be obtained, and the working waveforms can include voltage waveforms and current waveforms. Optionally, the voltage waveforms can be collected by a voltage sampling circuit, and the current waveforms can be collected by a current sampling circuit, but this is not limited. In addition, it should be noted that the voltage waveforms and the current waveforms included in each working waveform should have a corresponding relationship in time, that is, the start time of the voltage waveform should be the same as the start time of the current waveform, and the end time of the voltage waveform should be the same as the end time of the current waveform.

[0077] It can be understood that when obtaining the working waveforms of the first electrical apparatus in the plurality of opening and closing cycles, the plurality of working waveforms closest to the current time should be obtained as much as possible, so that when predicting the related parameters of the first electrical apparatus based on the plurality of working waveforms subsequently, a more accurate prediction result can be obtained. Of course, the number of working waveforms is not limited herein, and can be 50, 100, 200, or the like, which can be flexibly selected according to the actual application scenario.

[0078] S102, target feature data of each working waveform is extracted, and the target feature data includes wavelet packet features and / or arc parameters.

[0079] It can be understood that the target feature data can include wavelet packet features, or can include arc parameters, or can include wavelet packet features and arc parameters, which is not limited herein, and one or more feature data of each working waveform can be extracted as target feature data according to the actual application scenario. It can be understood that since the target feature data can be extracted from each working waveform, and each working waveform is the waveform of the first electrical apparatus in the plurality of opening and closing cycles within the historical time period, the extracted target feature data can represent the related working parameters of the first electrical apparatus, and thus when the target feature data is applied to the remaining life prediction of the first electrical apparatus, a more accurate remaining life parameter of the first electrical apparatus can be predicted. Of course, it should be noted that the extracted target feature data can also be used for other purposes, such as predicting the working state and usage frequency of the first electrical apparatus, which is not limited herein.

[0080] Of course, it should be noted that according to the actual application scenario, the target feature data further includes other feature data, such as contact voltage and contact current, which is not limited herein.

[0081] In summary, the embodiment of the present application provides a feature extraction method, comprising: obtaining working waveforms of a first electrical appliance in a plurality of opening and closing cycles in a historical time period, each working waveform comprising at least: a voltage waveform and a current waveform; extracting target feature data of each working waveform, the target feature data comprising: wavelet packet features, and / or arc parameters, by applying the embodiment of the present application, the extracted target feature data can represent the relevant working parameters of the first electrical appliance, and when the target feature data is applied to the remaining life prediction of the first electrical appliance, the remaining life parameters of the first electrical appliance can be predicted more accurately.

[0082] Figure 2 Another feature extraction method provided by the embodiment of the present application is shown in the flowchart. The extraction process of the target feature data is described below. As shown in Figure 2 The target feature data comprises: arc parameters, and the arc parameters comprise: arc energy proportion; the extraction of the target feature data of each working waveform comprises:

[0083] S201, according to the voltage waveform and the current waveform corresponding to each working waveform, the arc energy corresponding to each working waveform is calculated.

[0084] As can be seen from the definition of the arc energy, after obtaining the voltage waveform and the current waveform corresponding to each working waveform, the arc energy corresponding to each working waveform can be obtained by integrating the arc power in the arc time. Wherein, the calculation of the arc power can refer to the definition of the arc power, and the arc time can be determined according to the waveform start time and the waveform end time of the voltage waveform or the current waveform. Specifically, the arc time can be determined according to the time difference between the waveform end time and the waveform start time.

[0085] S202, based on the preset phase angle interval and the arc energy corresponding to each working waveform, a preset clustering algorithm is used to cluster each working waveform to obtain a clustering result.

[0086] As can be seen from the definition of the phase angle, the preset phase angle interval in the present application can be any phase angle interval between 0 and 360 degrees, such as 0 to 60 degrees, 60 to 150 degrees, 150 to 250 degrees, 250 to 360 degrees, etc., which is not limited here, and can be flexibly set according to the actual application scenario.

[0087] Optionally, when clustering, each working waveform can be clustered according to the preset clustering algorithm based on the preset phase angle interval and the arc energy corresponding to each working waveform. Of course, the application does not limit the category of clustering, and the category of clustering can be divided according to the arc energy. For example, the first category can be a working waveform with an arc energy between a first arc energy threshold and a second arc energy threshold; the second category can be a working waveform with an arc energy between the second arc energy threshold and a third arc energy threshold; and the third category can be a working waveform with an arc energy between the third arc energy threshold and a fourth arc energy threshold, wherein the fourth arc energy threshold is greater than the third arc energy threshold, the third arc energy threshold is greater than the second arc energy threshold, and the second arc energy threshold is greater than the first arc energy threshold, which is not limited herein.

[0088] S203, according to the clustering result, the arc energy proportion corresponding to the first electric appliance is calculated.

[0089] Based on the above description, it can be known that the clustering result can reflect the distribution of each working waveform in the preset phase angle interval, such as the clustering category and the clustering number. Therefore, the arc energy proportion corresponding to the first electric appliance can be calculated according to the clustering result, wherein the arc energy proportion corresponding to the first electric appliance can represent the ratio of the number of working waveforms meeting the preset arc energy threshold range to the total number of working waveforms.

[0090] Figure 3 Another feature extraction method provided by the embodiment of the application is shown in the flowchart as shown in Figure 3 According to the clustering result, the arc energy proportion corresponding to the first electric appliance is calculated, including:

[0091] S301, according to the clustering result, the number of working waveforms meeting the preset arc energy threshold range in the plurality of working waveforms is determined.

[0092] The preset arc energy threshold range can be a threshold range with lower arc energy in the clustering result. Based on the foregoing description, the preset arc energy threshold range can be a threshold range between the first arc energy threshold and the second arc energy threshold, but is not limited thereto.

[0093] It can be understood that, based on the above-determined clustering result, the number of working waveforms meeting the preset arc energy threshold range can be determined in the plurality of working waveforms. Optionally, the plurality of working waveforms can be 100, and the number of working waveforms meeting the preset arc energy threshold range can be 20, 30, or 50, which is not limited herein.

[0094] S302, according to the waveform quantity and the total quantity of the plurality of working waveforms, a first electric appliance corresponding arc energy proportion is calculated.

[0095] After obtaining the waveform quantity of the working waveforms meeting the preset arc energy threshold range, the first electric appliance corresponding arc energy proportion can be calculated according to the waveform quantity and the total quantity of the plurality of working waveforms. For example, the waveform quantity of the working waveforms meeting the preset arc energy threshold range in the plurality of working waveforms is 20, and the total quantity of the plurality of working waveforms is 100, and the arc energy proportion corresponding to the to-be-predicted electric appliance is 0.2. Of course, the actual calculation method is not limited to this.

[0096] Figure 4 A curve schematic diagram of the first electric appliance corresponding arc energy proportion and the number of operations of the embodiment of the present application is provided. As shown in the figure, when the first electric appliance corresponding arc energy proportion is used to represent the remaining life parameter of the first electric appliance, the horizontal axis represents the number of operations of the first electric appliance, and the vertical axis represents the first electric appliance corresponding arc energy proportion. Figure 4 As shown in the figure, the horizontal axis represents the number of operations of the first electric appliance, and the vertical axis represents the first electric appliance corresponding arc energy proportion. Figure 4 As a whole, it can be seen that the arc energy proportion will decrease with the increase of the number of operations of the first electric appliance, that is, the greater the arc energy proportion, the fewer the number of operations of the first electric appliance, that is, the more the remaining number of operations of the first electric appliance, and therefore, the first electric appliance corresponding arc energy proportion can reflect the remaining life parameter of the first electric appliance to a certain extent. Of course, the specific application of the first electric appliance corresponding arc energy proportion is not limited to this.

[0097] Figure 5 A flowchart of another feature extraction method provided by the embodiment of the present application is shown in the figure. Figure 6 A wavelet packet decomposition schematic diagram provided by the embodiment of the present application is shown in the figure. Optionally, the target feature data includes wavelet packet features, as shown in the figure. Figure 5 The above extraction of the target feature data of each working waveform includes:

[0098] S401, wavelet packet decomposition is performed on each working waveform to obtain a plurality of waveband waveforms corresponding to the decomposed working waveforms, wherein each waveband waveform corresponds to a different frequency interval.

[0099] When wavelet packet decomposition is performed, any number of layers of wavelet packet decomposition can be performed, such as 3 layers, 5 layers, 8 layers, etc., which is not limited here, and can be flexibly set according to the actual application scenario.

[0100] For example, as shown in the figure, Figure 6As shown, taking the three-layer wavelet packet decomposition of the voltage waveform in each working waveform as an example, specifically, in the wavelet decomposition, the voltage waveform can be filtered in the frequency domain through a group of high-pass filters and low-pass filters, the high-pass filters and low-pass filters can be called by a function library, and the input waveform is decomposed in the form of a binary tree, that is, one node in the binary tree can correspond to a wave segment waveform. According to the principle of wavelet packet decomposition, the voltage waveform can be decomposed into 8 wave segment waveforms through three-layer wavelet packet decomposition, which corresponds to the binary tree. Each wave segment waveform can be represented by (j, i), that is, the jth layer and the ith node. Referring to Figure 6 As shown, (1, 1), (1, 2), (2, 1), (2, 2), (2, 3), (2, 4), (3, 1), (3, 2), (3, 3), (3, 4), (3, 5), (3, 6), (3, 7), (3, 8) each node can represent each wave segment waveform, and each wave segment waveform can correspond to a different frequency interval. For example, the first wave segment waveform (1, 1) can correspond to the first frequency interval, the second wave segment waveform (1, 2) can correspond to the second frequency interval, the third wave segment waveform (2, 1) can correspond to the third frequency interval, and other wave segment waveforms can correspond to other frequency intervals. The present application will not be described here.

[0101] S402, calculate the frequency band energy of each wave segment waveform corresponding to each working waveform and the Shannon entropy of each working waveform.

[0102] After obtaining each wave segment waveform, the frequency band energy of each wave segment waveform can be further calculated. Alternatively, the frequency band energy can be calculated by the following formula:

[0103]

[0104] Wherein, E(j, i) can represent the frequency band energy of the ith node of the jth layer of decomposition, wherein when j = 1, the value of i can be 1 and 2, when j = 2, the value of i can be any integer in 1 to 4, and when j = 3, the value of i can be any integer in 1 to 8; p s (n, j, i) represents the wavelet packet transform coefficient of the jth node of the ith layer, which can be determined according to the high-pass filter and the low-pass filter; n represents n-layer decomposition, if it is three-layer wavelet packet decomposition, the value of n can be 3, after n-layer decomposition, the waveform to be decomposed can be decomposed into 2 n Waveforms, wherein each wave segment waveform can correspond to a corresponding frequency band signal.

[0105] It should be noted that when performing the calculation, since each working waveform includes a voltage waveform and a current waveform, the band energy and the Shannon entropy of the voltage waveform and the current waveform in each working waveform can be calculated respectively when performing the calculation, that is, for any voltage waveform, the band energy of each band waveform corresponding to the voltage waveform and the Shannon entropy corresponding to the voltage waveform can be calculated; for any current waveform, the band energy of each band waveform corresponding to the current waveform and the Shannon entropy corresponding to the current waveform can be calculated. Wherein, for each working waveform, the band energy of each band waveform corresponding to the voltage waveform, the Shannon entropy corresponding to the voltage waveform, the band energy of each band waveform corresponding to the current waveform and the Shannon entropy corresponding to the current waveform can be taken as the target feature data of each working waveform.

[0106] Based on the above description, after calculating the band energy of each band waveform corresponding to each working waveform, the Shannon entropy of each working waveform can be further calculated according to the band energy of each band waveform in each working waveform, that is, the Shannon entropy corresponding to the voltage waveform and the current waveform in each working waveform can be calculated respectively.

[0107] Figure 7 A flowchart of another feature extraction method provided by the embodiments of the present application is shown in FIG. 6. Optionally, as shown in FIG. 6, the above-mentioned calculation of the Shannon entropy of each working waveform includes: Figure 7

[0108] S501, calculating the normalized band energy coefficient of each band waveform in each working waveform according to the band energy of each band waveform corresponding to each working waveform.

[0109] S502, calculating the Shannon entropy of each working waveform according to the normalized band energy coefficient of each band waveform in each working waveform.

[0110] Wherein, the normalized band energy coefficient of each band waveform in the voltage waveform or the current waveform in the working waveform can be calculated according to the ratio of the band energy of each band waveform to the sum of the band energy of each band waveform, and the sum of the band energy of each band waveform can be calculated by summing the band energy of each band waveform.

[0111] Optionally, the process of calculating the Shannon entropy of each working waveform according to the normalized band energy coefficient of each band waveform in each working waveform can be calculated according to the following formula:

[0112]

[0113] Wherein, x w (t) represents the voltage waveform or the current waveform in the working waveform, represents the calculation of x w ​(t) is a normalized band energy coefficient of the Nth band waveform after n-layer decomposition, n represents n-layer decomposition, if it is three-layer wavelet packet decomposition, n can be 3, after n-layer decomposition, the waveform to be decomposed can be decomposed into 2 n band waveforms, H(x w (t)) represents the Shannon entropy of the voltage waveform or the current waveform x w (t).

[0114] For example, based on the above examples, after three-layer wavelet packet decomposition of the voltage waveform, eight band waveforms can be obtained, and the normalized band energy coefficient of the first band waveform can be calculated according to the ratio of the band energy of the first band waveform to the sum of the band energies of the eight band waveforms. The normalized band energy coefficients of other band waveforms of the voltage waveform can be calculated according to the normalized band energy coefficient of the first band waveform, which will not be repeated here. In summary, it can be understood that if three-layer wavelet packet decomposition is performed, nine feature values can be used to represent a voltage waveform, including eight band energy features and one Shannon entropy feature, so as to reflect the characteristics of the voltage waveform in as many dimensions as possible.

[0115] Optionally, the preset clustering algorithm can include at least one of the following: K-means clustering algorithm, mean shift clustering algorithm, hierarchical clustering algorithm. In some embodiments, the preset clustering algorithm can also be based on density algorithms (such as DBSCAN), graph clustering methods, grid algorithms, etc., which are not limited here.

[0116] Optionally, the target feature data further includes at least one of the following: arcing time, arcing power, contact voltage, contact current, contact resistance. The explanations of these features can be found in the aforementioned glossary, which will not be repeated here. It should be noted that, according to the actual application scenario and the specific category of the first electrical appliance, the target feature data can also include other categories of feature data, which are not limited here.

[0117] Optionally, the present embodiment further provides a low-voltage electrical appliance life prediction method, comprising:

[0118] Inputting the target feature data corresponding to the electrical appliance to be predicted into the life prediction model to predict and obtain the remaining life parameter of the electrical appliance to be predicted.

[0119] The target feature data corresponding to the to-be-predicted electrical appliance is obtained by using the foregoing feature extraction method, and the first electrical appliance is the to-be-predicted electrical appliance. The life prediction model is obtained by training based on a training sample data set. The training sample data set includes target feature data corresponding to a plurality of sample electrical appliances and labeled corresponding residual life parameters. The target feature data corresponding to the plurality of sample electrical appliances is obtained by using the foregoing feature extraction method, and the first electrical appliance is the plurality of sample electrical appliances.

[0120] Optionally, the life prediction model can be implemented based on a deep learning technology, a random forest technology, etc., which is not limited herein. Optionally, the electrical appliance types of the plurality of sample electrical appliances can be the same as the electrical appliance type of the to-be-predicted electrical appliance, for example, can all be contactors, or can all be relays, so that when the life prediction model obtained based on the target feature data corresponding to the plurality of sample electrical appliances is used to predict the residual life parameter of the to-be-predicted electrical appliance, a relatively accurate prediction result can be obtained. It can be understood that for the plurality of sample electrical appliances, that is, the working waveforms of the plurality of sample electrical appliances in the plurality of opening and closing cycles in the historical time period can be obtained by using the foregoing feature extraction method. Each working waveform at least includes a voltage waveform and a current waveform. The target feature data of each working waveform is extracted, so that the target feature data corresponding to the plurality of sample electrical appliances can be extracted.

[0121] In the specific training of the life prediction model, the extracted target feature data corresponding to the plurality of sample electrical appliances can be used as a training sample data set, each target feature data is labeled with a corresponding residual life parameter, and the life prediction model is obtained by training. The obtained life prediction model can be used to predict the residual life parameter of the to-be-predicted electrical appliance.

[0122] Optionally, the residual life parameter can be a residual working time, a residual working frequency, etc., which is not limited herein and can be different according to different sample electrical appliances and to-be-predicted electrical appliances. In addition, it can be understood that since the life prediction model is obtained by training based on the target feature data corresponding to the plurality of sample electrical appliances, the life prediction model obtained by training can learn the relationship between the target feature data corresponding to the plurality of sample electrical appliances and the residual life parameter, and when the life prediction model is used to predict the residual life parameter of the to-be-predicted electrical appliance, a relatively accurate life prediction result can be obtained. In addition, it can be understood that the categories of the target feature data corresponding to the plurality of sample electrical appliances should be the same as the categories of the target feature data corresponding to the to-be-predicted electrical appliance, that is, if the target feature data corresponding to the plurality of sample electrical appliances further includes at least one of the following feature data: arcing time, arcing power, contact voltage, contact current, contact resistance, or other categories of feature data, then the target feature data corresponding to the to-be-predicted electrical appliance should also include the corresponding feature data.

[0123] It should be further explained that according to actual application scenarios, if there is a higher requirement for the prediction time (for example, the prediction time is required to be as short as possible), any one of the wavelet packet feature and the arcing parameter can be selected as the target feature data corresponding to the plurality of sample data to reduce the time required for extracting the target feature data; and if there is a higher requirement for the prediction accuracy, the wavelet packet feature and the arcing parameter can be extracted as the target feature data corresponding to the plurality of sample data, or other features can be extracted as the target feature data corresponding to the plurality of sample data to enrich the types of the target feature data, which is not limited herein.

[0124] In summary, the application provides a low-voltage electrical appliance life prediction method, including the feature extraction method in any of the preceding embodiments, wherein the first electrical appliance includes a to-be-predicted electrical appliance and a plurality of sample electrical appliances, and the low-voltage electrical appliance life prediction method includes: inputting the target feature data corresponding to the to-be-predicted electrical appliance into a life prediction model to predict and obtain a remaining life parameter of the to-be-predicted electrical appliance, wherein the life prediction model is trained according to a training sample data set, and the training sample data set includes the target feature data corresponding to the plurality of sample electrical appliances and the corresponding labeled remaining life parameter. By applying the embodiment of the application, the remaining life parameter of the to-be-predicted electrical appliance can be predicted and obtained through the life prediction model. Compared with the prior art method of predicting the remaining life according to the installation time of the to-be-predicted electrical appliance, the embodiment of the application can obtain an accurate life prediction result.

[0125] Figure 8 A functional module schematic diagram of a feature extraction device provided by the embodiment of the application is shown in the figure. The basic principle and technical effects of the device are the same as those of the corresponding method embodiments described above. For brief description, the parts not mentioned in the embodiment can refer to the corresponding contents in the method embodiments. As shown in the figure, the feature extraction device 100 includes: Figure 8

[0126] The acquisition module 110 is configured to acquire working waveforms of the first electrical appliance in a plurality of opening and closing cycles in a historical time period, and each working waveform includes at least a voltage waveform and a current waveform;

[0127] The extraction module 120 is configured to extract target feature data of each working waveform, and the target feature data includes wavelet packet features and / or arcing parameters.

[0128] In an optional implementation, the target feature data includes arcing parameters, and the arcing parameters include an arcing energy proportion. The extraction module 120 is specifically configured to calculate arcing energy corresponding to each working waveform according to the voltage waveform and the current waveform corresponding to each working waveform; perform clustering on each working waveform based on a preset phase angle interval and the arcing energy corresponding to each working waveform by using a preset clustering algorithm to obtain a clustering result; and calculate the arcing energy proportion corresponding to the first electrical appliance according to the clustering result.​

[0129] In an optional implementation, the extraction module 120 is specifically configured to determine, according to the clustering result, a waveform quantity of working waveforms in the plurality of working waveforms that meet a preset arcing energy threshold range; and calculate an arcing energy proportion of the first electrical appliance according to the waveform quantity and a total quantity of the plurality of working waveforms.

[0130] In an optional implementation, the target feature data includes wavelet packet features, and the extraction module 120 is specifically configured to perform wavelet packet decomposition on each working waveform to obtain a plurality of waveband waveforms corresponding to the working waveform after decomposition, wherein each waveband waveform corresponds to a different frequency interval; and calculate a frequency band energy of each waveband waveform corresponding to each working waveform and a Shannon entropy of each working waveform.

[0131] In an optional implementation, the extraction module 120 is specifically configured to calculate, according to the frequency band energy of each waveband waveform corresponding to each working waveform, a normalized frequency band energy coefficient of each waveband waveform in each working waveform; and calculate a Shannon entropy of each working waveform according to the normalized frequency band energy coefficient of each waveband waveform in each working waveform.

[0132] In an optional implementation, the preset clustering algorithm includes at least one of the following: a K-means clustering algorithm, a mean shift clustering algorithm, and a hierarchical clustering algorithm.

[0133] In an optional implementation, the target feature data further includes at least one of the following: arcing time, arcing power, contact voltage, contact current, and contact resistance.

[0134] Optionally, the embodiments of the present application also provide a low-voltage electrical appliance life prediction device, and the basic principles and technical effects generated by the device are the same as those of the corresponding method embodiments described above. For brevity, the parts not mentioned in the present embodiment can be referred to the corresponding contents in the method embodiments. The low-voltage electrical appliance life prediction device can include:

[0135] The prediction module is configured to input target feature data corresponding to a to-be-predicted electrical appliance into a life prediction model to predict and obtain a remaining life parameter of the to-be-predicted electrical appliance, wherein the target feature data corresponding to the to-be-predicted electrical appliance is obtained by using the feature extraction method described above, and the first electrical appliance is the to-be-predicted electrical appliance.

[0136] The life prediction model is trained and obtained according to a training sample data set, the training sample data set includes target feature data corresponding to a plurality of sample electrical appliances and labels corresponding remaining life parameters, the target feature data corresponding to the plurality of sample electrical appliances is obtained by using the feature extraction method described above, and the first electrical appliance is the plurality of sample electrical appliances.

[0137] The above device is used to execute the method provided by the above embodiments, and the implementation principle and technical effects are similar, which will not be described here.

[0138] The above modules can be one or more integrated circuits configured to implement the above methods, for example, one or more application specific integrated circuits (ASICs), or one or more microprocessors, or one or more field programmable gate arrays (FPGAs), etc. For another example, when a certain module above is implemented in the form of a processing element scheduling code, the processing element can be a general purpose processor, such as a central processing unit (CPU) or other processor that can invoke program code. For another example, the modules can be integrated together to implement a system-on-a-chip (SOC).

[0139] Figure 9 An electronic device structure schematic diagram is provided for an embodiment of the present application. The electronic device can be integrated in the first appliance described above. As shown in the figure, the electronic device can include a processor 210, a storage medium 220, and a bus 230. The storage medium 220 stores machine readable instructions executable by the processor 210. When the electronic device is running, the processor 210 communicates with the storage medium 220 through the bus 230. The processor 210 executes the machine readable instructions to perform the steps of the method embodiments described above. The specific implementation and technical effects are similar, and will not be repeated here. Figure 9

[0140] Optionally, the present application also provides a storage medium, and the storage medium stores a computer program. When the computer program is run by a processor, the steps of the method embodiments described above are performed. The specific implementation and technical effects are similar, and will not be repeated here.

[0141] In several embodiments provided by the present application, it should be understood that the disclosed apparatus and method can be implemented in other ways. For example, the apparatus embodiments described above are merely schematic. For example, the division of units is merely a logical function division. In actual implementation, another division manner can be used. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units or components shown or discussed can be indirect coupling or communication connection through some interfaces. The coupling or communication connection can be electrical, mechanical or in other forms.

[0142] ​The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, i.e. may be located in one place, or may be distributed on multiple network units. Part or all of the units may be selected according to actual needs to achieve the purpose of the embodiment.

[0143] In addition, the functional units in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware, or in the form of hardware plus software function unit.

[0144] The integrated unit realized in the form of software function unit can be stored in a computer readable storage medium. The software function unit stored in a storage medium includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor (English: processor) to execute part of the steps of the method of each embodiment of the present application. The storage medium mentioned above includes: U disk, mobile hard disk, read-only memory (English: Read-Only Memory, for short: ROM), random access memory (English: Random Access Memory, for short: RAM), magnetic disk or optical disk and various program code storage media.

[0145] It should be noted that in this paper, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between the entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment. Without more limitations, the element defined by the statement "including a" does not exclude the presence of other identical elements in the process, method, article or equipment including the element.

[0146] The above merely provides preferred embodiments of the present application, and is not intended to limit the present application. Based on the embodiments of the present application, those of ordinary skill in the art can make various modifications and variations without departing from the spirit and principle of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall fall within the scope of the present application. It should be noted that similar reference numerals and letters represent similar items in the following drawings, and thus, once an item is defined in one drawing, it need not be further defined and explained in subsequent drawings. The above merely provides preferred embodiments of the present application, and is not intended to limit the present application. Based on the embodiments of the present application, those of ordinary skill in the art can make various modifications and variations without departing from the spirit and principle of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall fall within the scope of the present application.

Claims

1. A feature extraction method characterized by, The method comprises: acquiring working waveforms of the first electrical appliance in a plurality of opening and closing cycles in a historical time period, each of the working waveforms comprising at least: a voltage waveform and a current waveform; extracting target feature data of each of the working waveforms, the target feature data comprising: wavelet packet features and / or arc parameters; the target feature data comprises arc parameters, the arc parameters comprising: an arc energy proportion; the extracting of the target feature data of each of the working waveforms comprises: calculating arc energies corresponding to each of the working waveforms according to the voltage waveforms and the current waveforms corresponding to each of the working waveforms; clustering each of the working waveforms based on preset phase angle intervals and the arc energies corresponding to each of the working waveforms by using a preset clustering algorithm to obtain a clustering result; and calculating the arc energy proportion corresponding to the first electrical appliance according to the clustering result; the calculating of the arc energy proportion corresponding to the first electrical appliance according to the clustering result comprises: determining a waveform number of working waveforms in the plurality of working waveforms that meet a preset arc energy threshold range according to the clustering result; and calculating the arc energy proportion corresponding to the first electrical appliance according to the waveform number and a total number of the plurality of working waveforms; the target feature data comprises wavelet packet features, and the extracting of the target feature data of each of the working waveforms comprises: performing wavelet packet decomposition on each of the working waveforms to obtain a plurality of waveband waveforms corresponding to each of the working waveforms after decomposition, wherein each of the waveband waveforms corresponds to a different frequency interval; and calculating frequency band energies of each of the waveband waveforms corresponding to each of the working waveforms and a Shannon entropy of each of the working waveforms.

2. The method of claim 1, wherein, the calculating of the Shannon entropy of each of the working waveforms comprises: calculating a normalized frequency band energy coefficient of each of the waveband waveforms in each of the working waveforms according to the frequency band energies of each of the waveband waveforms corresponding to each of the working waveforms; and calculating the Shannon entropy of each of the working waveforms according to the normalized frequency band energy coefficients of each of the waveband waveforms in each of the working waveforms.

3. The method of claim 1, wherein, The preset clustering algorithm comprises at least one of the following: a K-means clustering algorithm, a mean shift clustering algorithm, and a hierarchical clustering algorithm.

4. The method according to any one of claims 1 to 3, characterized in that, The target feature data further comprises at least one of the following: an arc time, an arc power, a contact voltage, a contact current, and a contact resistance.

5. A low voltage electrical apparatus life prediction method, characterized by, The method comprises: inputting target feature data corresponding to a to-be-predicted electrical appliance into a life prediction model to predict and acquire a remaining life parameter of the to-be-predicted electrical appliance, wherein the target feature data corresponding to the to-be-predicted electrical appliance is obtained by using any one of the feature extraction methods in claims 1-4, the first electrical appliance is the to-be-predicted electrical appliance, the life prediction model is obtained by training according to a training sample data set, the training sample data set comprises target feature data corresponding to a plurality of sample electrical appliances and labels corresponding remaining life parameters, the target feature data corresponding to the plurality of sample electrical appliances is obtained by using any one of the feature extraction methods in claims 1-4, and the first electrical appliance is the plurality of sample electrical appliances.

6. A feature extraction apparatus characterized by comprising: The method comprises: An acquisition module is configured to acquire working waveforms of the first electric appliance in a plurality of opening and closing cycles in a historical time period, each of the working waveforms comprising at least a voltage waveform and a current waveform; An extraction module is configured to extract target feature data of each of the working waveforms, the target feature data comprising wavelet packet features and / or arc parameters; The target feature data comprises arc parameters, the arc parameters comprising an arc energy proportion, and the extraction module is specifically configured to calculate arc energy corresponding to each of the working waveforms according to the voltage waveform and the current waveform corresponding to each of the working waveforms; perform clustering on each of the working waveforms based on a preset phase angle interval and the arc energy corresponding to each of the working waveforms by using a preset clustering algorithm to obtain a clustering result; and calculate the arc energy proportion corresponding to the first electric appliance according to the clustering result. The calculation of the arc energy proportion corresponding to the first electric appliance according to the clustering result comprises: determining a waveform number of working waveforms in the plurality of working waveforms that meet a preset arc energy threshold range according to the clustering result; and calculating the arc energy proportion corresponding to the first electric appliance according to the waveform number and a total number of the plurality of working waveforms. The target feature data comprises wavelet packet features, and the extraction module is specifically configured to perform wavelet packet decomposition on each of the working waveforms to obtain a plurality of waveband waveforms corresponding to each of the working waveforms after decomposition, wherein each of the waveband waveforms corresponds to a different frequency interval; and calculate frequency band energy of each waveband waveform corresponding to each of the working waveforms and Shannon entropy of each of the working waveforms.

7. A computer readable storage medium characterized in that, The computer readable storage medium stores a computer program, and the computer program is run by the processor to execute the steps of the feature extraction method according to any one of claims 1-4 or the steps of the low-voltage electric appliance life prediction method according to claim 5.

Images