Low-voltage cable fault identification method and device, computer equipment and storage medium

Abstract

The application relates to a low-voltage cable fault identification method and device, computer equipment and a storage medium, which are applied to the field of fault diagnosis. The method comprises the following steps: acquiring a frequency domain waveform corresponding to a to-be-identified low-voltage cable; the frequency domain waveform is obtained by testing the to-be-identified low-voltage cable based on a frequency domain reflection method; performing Fourier transform processing on the frequency domain waveform to obtain a distance diagnosis graph corresponding to the to-be-identified low-voltage cable; inputting the distance diagnosis graph into a trained low-voltage cable fault identification model to obtain a fault type corresponding to the to-be-identified low-voltage cable. The method can improve the identification efficiency of the low-voltage cable and the accuracy of fault diagnosis.

Description

Technical Field

[0001] This application relates to the field of fault diagnosis technology, and in particular to a method, apparatus, computer equipment, storage medium and computer program product for identifying faults in low-voltage cables. Background Technology

[0002] The low-voltage system of a substation is responsible for providing power to the secondary equipment within the substation. The stability of the low-voltage system directly affects the reliability and stability of the substation's operation. During operation, the cables used in the substation's low-voltage system are subjected to factors such as thermal stress, electrical stress, and external damage, causing a gradual decline in insulation performance and posing potential hazards to the safe operation of the substation. If local defects such as damage to the outer sheath or moisture intrusion occur inside the low-voltage cable, it can lead to faults such as arcing and grounding in the substation's AC system. This can result in minor energy losses, or even serious accidents such as malfunctions of the substation's protection system or fires, posing a serious threat to the power system. If the type of cable defect in the low-voltage system can be accurately determined in a timely manner, and the cable insulation performance can be assessed, targeted maintenance and repair of the substation's low-voltage system can be carried out, further eliminating potential hazards caused by low-voltage cable faults. Therefore, strengthening the monitoring of the insulation status of cables in the substation's low-voltage system and improving the technology for identifying local defects in low-voltage cables are of great significance for the safe operation of the power system. Summary of the Invention

[0003] Therefore, it is necessary to provide a low-voltage cable fault identification method, device, computer equipment, computer-readable storage medium, and computer program product to address the above problems.

[0004] Firstly, this application provides a method for identifying faults in low-voltage cables. The method includes:

[0005] Obtain the frequency domain waveform corresponding to the low-voltage cable to be identified; the frequency domain waveform is obtained by testing the low-voltage cable to be identified based on the frequency domain reflection method;

[0006] The frequency domain waveform is subjected to Fourier transform processing to obtain the distance diagnostic spectrum corresponding to the low-voltage cable to be identified;

[0007] The distance diagnostic map is input into the trained low-voltage cable fault identification model to obtain the fault type corresponding to the low-voltage cable to be identified.

[0008] In one embodiment, performing Fourier transform processing on the frequency domain waveform to obtain the distance diagnostic map corresponding to the low-voltage cable to be identified includes:

[0009] The reflection coefficient in the spectrum corresponding to the frequency domain waveform is transformed into a time domain signal that varies with time.

[0010] Perform a Fast Fourier Transform or Discrete Fourier Transform on the real or imaginary part of the waveform spectrum of the transformed time-domain signal to obtain the reflection coefficient spectrum that varies with the fundamental frequency.

[0011] The frequency coordinates in the reflection coefficient spectrum are converted into cable distance coordinates to obtain a distance diagnostic spectrum of reflection coefficient as a function of distance, which is used as the distance diagnostic spectrum of the low-voltage cable to be identified.

[0012] In one embodiment, converting the frequency coordinates in the reflection coefficient spectrum into cable distance coordinates to obtain a distance diagnostic spectrum of reflection coefficient as a function of distance includes:

[0013] The frequency coordinates in the reflection coefficient spectrum are converted into cable distance coordinates to obtain the original distance diagnostic spectrum of reflection coefficient as a function of distance;

[0014] The original distance diagnostic map is processed by adding a distance window to obtain the distance diagnostic map in which the reflectance coefficient changes with distance.

[0015] In one embodiment, the low-voltage cable fault identification model is trained in the following manner:

[0016] Test the low-voltage cables with known fault types using the frequency domain reflection method to obtain the corresponding frequency domain waveforms;

[0017] Perform a Fourier transform on the frequency domain waveform of the low-voltage cable with the known fault type to obtain the distance diagnostic spectrum corresponding to the low-voltage cable with the known fault type;

[0018] Input the distance diagnostic map corresponding to the low-voltage cable with the known fault type into the low-voltage cable fault identification model to be trained to obtain the predicted fault type;

[0019] Based on the difference information between the predicted fault type and the known fault type, the low-voltage cable fault identification model to be trained is trained to obtain the trained low-voltage cable fault identification model.

[0020] In one embodiment, the test of sample low-voltage cables with known fault types based on the frequency domain reflection method includes:

[0021] A rectangular insulation layer of a set area is stripped from phase A of the exposed insulation portion of a normal low-voltage cable to expose the cable core, forming a first defect. At the first defect, a test based on the frequency domain reflection method is performed between the cable core and the copper shielding layer with resistors of different resistance values ​​to obtain the frequency domain waveform corresponding to phase A.

[0022] A longitudinal knife mark defect of a set length and depth is made on phase B at the first distance from the cable head of the exposed insulation part of a normal low-voltage cable to form a second defect. At the second defect, a test based on the frequency domain reflection method is performed between the cable core and the copper shielding layer with resistors of different resistance values ​​to obtain the frequency domain waveform corresponding to phase B.

[0023] A V-shaped defect is made along the radial direction of the cable on phase C at the second distance from the cable head end of the exposed insulation part of the normal low-voltage cable, forming a third defect. At the third defect, a test based on the frequency domain reflection method is performed between the cable core and the copper shielding layer with resistors of different resistance values ​​to obtain the frequency domain waveform corresponding to C.

[0024] In one embodiment, before inputting the distance diagnostic map corresponding to the low-voltage cable with the known fault type into the low-voltage cable fault identification model to be trained to obtain the predicted fault type, the method further includes:

[0025] The distance diagnostic map of the low-voltage cable with the known fault type is deformed to obtain a data-enhanced distance diagnostic map; the deformation process includes at least one of trimming, compression and rotation;

[0026] The step of inputting the distance diagnostic map corresponding to the low-voltage cable with the known fault type into the low-voltage cable fault identification model to be trained to obtain the predicted fault type includes:

[0027] The augmented distance diagnostic map is input into the low-voltage cable fault identification model to be trained to obtain the predicted fault type.

[0028] Secondly, this application also provides a low-voltage cable fault identification device. The device includes:

[0029] The acquisition module is used to acquire the frequency domain waveform corresponding to the low-voltage cable to be identified; the frequency domain waveform is obtained by testing the low-voltage cable to be identified based on the frequency domain reflection method;

[0030] The transformation module is used to perform Fourier transform processing on the frequency domain waveform to obtain the distance diagnostic spectrum corresponding to the low-voltage cable to be identified;

[0031] The identification module is used to input the distance diagnostic map into the trained low-voltage cable fault identification model to obtain the fault type corresponding to the low-voltage cable to be identified.

[0032] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to perform the following steps:

[0033] Obtain the frequency domain waveform corresponding to the low-voltage cable to be identified; the frequency domain waveform is obtained by testing the low-voltage cable to be identified based on the frequency domain reflection method;

[0034] The frequency domain waveform is subjected to Fourier transform processing to obtain the distance diagnostic spectrum corresponding to the low-voltage cable to be identified;

[0035] The distance diagnostic map is input into the trained low-voltage cable fault identification model to obtain the fault type corresponding to the low-voltage cable to be identified.

[0036] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, performs the following steps:

[0037] Obtain the frequency domain waveform corresponding to the low-voltage cable to be identified; the frequency domain waveform is obtained by testing the low-voltage cable to be identified based on the frequency domain reflection method;

[0038] The frequency domain waveform is subjected to Fourier transform processing to obtain the distance diagnostic spectrum corresponding to the low-voltage cable to be identified;

[0039] The distance diagnostic map is input into the trained low-voltage cable fault identification model to obtain the fault type corresponding to the low-voltage cable to be identified.

[0040] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, performs the following steps:

[0041] Obtain the frequency domain waveform corresponding to the low-voltage cable to be identified; the frequency domain waveform is obtained by testing the low-voltage cable to be identified based on the frequency domain reflection method;

[0042] The frequency domain waveform is subjected to Fourier transform processing to obtain the distance diagnostic spectrum corresponding to the low-voltage cable to be identified;

[0043] The distance diagnostic map is input into the trained low-voltage cable fault identification model to obtain the fault type corresponding to the low-voltage cable to be identified.

[0044] The aforementioned low-voltage cable fault identification method, device, computer equipment, storage medium, and computer program product are based on the frequency domain reflection method. They utilize the characteristics of abundant injected high-frequency signals, higher signal-to-noise ratio, and the ability to identify weak defects to conduct frequency domain reflection tests on the low-voltage cables to be identified. Based on the frequency domain waveform characteristics obtained from the tests, local defects in the cables can be identified. This can effectively reduce the potential for power supply accidents caused by local faults in low-voltage cables in substations. By training a low-voltage cable fault identification model, the efficiency of low-voltage cable identification and the accuracy of fault diagnosis can be improved. Attached Figure Description

[0045] Figure 1 This is an application environment diagram of a low-voltage cable fault identification method in one embodiment;

[0046] Figure 2 This is a flowchart illustrating a low-voltage cable fault identification method in one embodiment;

[0047] Figure 3 This is a flowchart illustrating the training steps of a low-voltage cable fault identification model in one embodiment.

[0048] Figure 4 This is a schematic diagram illustrating a test performed on a defective cable sample using the frequency domain reflection method in one embodiment.

[0049] Figure 5 This is a schematic diagram of a defective cable sample in one embodiment;

[0050] Figure 6 This is a schematic diagram of the frequency domain waveforms obtained by connecting resistors of different resistance values ​​to a defect in phase D1 of a low-voltage cable in one embodiment.

[0051] Figure 7 A schematic diagram of the frequency domain waveforms before and after creating the D2 and D3 tool mark defects in phases B and C in one embodiment;

[0052] Figure 8 This is a structural block diagram of a low-voltage cable fault identification device in one embodiment;

[0053] Figure 9 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0055] The low-voltage cable fault identification method provided in this application can be applied to, for example... Figure 1In the application environment shown, test device 102 is communicatively connected to terminal 104. Test device 102 can be a vector network analyzer. Terminal 104 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, etc. Portable wearable devices can include smartwatches, smart bracelets, head-mounted devices, etc.

[0056] In the application scenario of this application, the test device 102 performs a test on the low-voltage cable to be identified based on the frequency domain reflection method to obtain the corresponding frequency domain waveform. The terminal 104 obtains the corresponding frequency domain waveform from the test device 102 and performs Fourier transform processing on the frequency domain waveform to obtain the distance diagnostic map corresponding to the low-voltage cable to be identified. The distance diagnostic map is input into the trained low-voltage cable fault identification model to obtain the fault type corresponding to the low-voltage cable to be identified.

[0057] In one embodiment, such as Figure 2 As shown, a low-voltage cable fault identification method is provided, which can be applied to... Figure 1 Taking terminal 104 as an example, the explanation includes the following steps:

[0058] Step S210: Obtain the frequency domain waveform corresponding to the low-voltage cable to be identified; the frequency domain waveform is obtained by testing the low-voltage cable to be identified based on the frequency domain reflection method.

[0059] Specifically, before performing a frequency domain reflection (FDR) test on the low-voltage cable to be identified, the cable core at the first end (test end) of the low-voltage cable sample is connected to the test lead of the frequency modulation signal source, the copper shielding layer of the sample is grounded, and the end of the sample is open-circuited. Then, the terminal (PC) controls the output frequency (0.15-200MHz) of the frequency modulation signal source, and applies a frequency modulation voltage to the low-voltage cable to be identified through the modulation signal source to perform a frequency domain reflection test. The reflected signal is received by the test equipment 102 to obtain the frequency domain waveform of the low-voltage cable to be identified.

[0060] A modulated signal V is transmitted to the beginning of the low-voltage cable to be identified via an FM signal source. i (f) The reflected signal V is measured by the test equipment 102. r (f), based on the transmitted modulation signal V i (f) and reflected signal V r (f) Obtain the reflection coefficient f represents the test signal frequency. Based on the obtained reflection coefficient Г(f), the frequency domain waveform spectrum of the low-voltage cable to be identified is obtained.

[0061] Step S220: Perform Fourier transform processing on the frequency domain waveform to obtain the distance diagnostic spectrum corresponding to the low-voltage cable to be identified.

[0062] In practice, performing a Fourier transform on a frequency domain waveform involves first transforming the reflection coefficient in the spectrum corresponding to the frequency domain waveform into a time-domain signal that varies with time. Then, a Fast Fourier Transform or Discrete Fourier Transform is performed on the real or imaginary part of the waveform spectrum of the transformed time-domain signal to obtain a reflection coefficient spectrum that varies with the fundamental frequency. Further, the frequency coordinates in the reflection coefficient spectrum are converted into cable distance coordinates to obtain the original distance diagnosis spectrum that varies with distance. Finally, a distance window is applied to the original distance diagnosis spectrum to obtain the distance diagnosis spectrum.

[0063] Step S230: Input the distance diagnostic map into the trained low-voltage cable fault identification model to obtain the fault type corresponding to the low-voltage cable to be identified.

[0064] The fault types can include normal, high resistance fault, low resistance fault, and tool mark fault.

[0065] In the specific implementation, a sample training dataset can be pre-acquired, which includes sample low-voltage cables and their fault type labels. Using the distance diagnostic map of the sample low-voltage cables as input variables, the low-voltage cable fault identification model to be trained is input to obtain the predicted fault type of the sample low-voltage cables. The loss value between the predicted fault type and the fault type label is calculated. Training terminates when the loss value is less than a preset threshold or when a preset number of training iterations are reached, resulting in a trained low-voltage cable fault identification model. Furthermore, after obtaining the distance diagnostic map of the low-voltage cable to be identified, the distance diagnostic map can be input into the trained low-voltage cable fault identification model to obtain the corresponding fault type of the low-voltage cable to be identified.

[0066] In an exemplary embodiment, the low-voltage cable fault identification model constructed in this application specifically includes: an input layer, a first convolutional layer, a first pooling layer, a second convolutional layer, a second pooling layer, a third convolutional layer, and a third pooling layer cascaded in sequence. The outputs of the third convolutional layer and the third pooling layer are both connected to a fourth convolutional layer, and the fourth convolutional layer is connected to a Softmax layer.

[0067] In the aforementioned low-voltage cable fault identification method, the frequency domain reflection method is used. This method leverages the richness of injected high-frequency signals, higher signal-to-noise ratio, and the ability to identify weak defects. The low-voltage cable to be identified is tested using the frequency domain reflection method, and the resulting frequency domain waveform characteristics are used to identify local defects in the cable. This effectively reduces the potential for power supply accidents caused by local faults in low-voltage cables within substations. Training a low-voltage cable fault identification model to identify local defects in low-voltage cables can improve the efficiency of identification and the accuracy of fault diagnosis. Furthermore, the local defect diagnosis technology based on the frequency domain reflection method proposed in this application is feasible for low-voltage PVC cables. This method is characterized by simple testing and high sensitivity. Moreover, the frequency domain reflection method is not limited by cable insulation materials or cable laying methods, thus enabling the diagnosis of local defects in various cable types.

[0068] In an exemplary embodiment, step S220 above involves performing a Fourier transform on the frequency domain waveform to obtain a distance diagnostic map corresponding to the low-voltage cable to be identified, including:

[0069] Step S221: Transform the reflection coefficient in the spectrum corresponding to the frequency domain waveform into a time domain signal that varies with time.

[0070] Step S222: Perform a fast Fourier transform or a discrete Fourier transform on the real or imaginary part of the waveform spectrum of the transformed time-domain signal to obtain a reflection coefficient spectrum that shows the reflection coefficient as a function of the fundamental frequency.

[0071] Step S223: Convert the frequency coordinates in the reflection coefficient spectrum into cable distance coordinates to obtain a distance diagnostic spectrum of the reflection coefficient as a function of distance, which is used as the distance diagnostic spectrum of the low-voltage cable to be identified.

[0072] In specific implementation, the reflection coefficient in the frequency domain waveform of the low-voltage cable to be identified is transformed into a time domain signal that varies with time t′. A Fast Fourier Transform or Discrete Fourier Transform is performed on the real part Real(Г(t′)) or imaginary part Imag(Г(t′)) of the transformed waveform spectrum to obtain a reflection coefficient spectrum that varies with the fundamental frequency f′. The frequency coordinates are then converted to cable distance coordinates to obtain a distance diagnostic spectrum D(i) that shows the reflection coefficient varying with distance. The conversion method for converting the frequency coordinates f′ to the cable distance coordinates l is as follows:

[0073]

[0074] Where l represents the cable distance coordinate, and f′ represents the frequency coordinate. c is the speed of light, ε r is the relative permittivity.

[0075] Furthermore, in an exemplary embodiment, in order to increase the sensitivity of local defect identification, this application also proposes a method for adding a distance window to the original distance diagnostic map D0. The above step S223 further includes: converting the frequency coordinates in the reflection coefficient map into cable distance coordinates to obtain the original distance diagnostic map of the reflection coefficient changing with distance; adding a distance window to the original distance diagnostic map to obtain the distance diagnostic map of the reflection coefficient changing with distance.

[0076] In practice, the boundary range of the window length of the original distance diagnostic map can be determined first. Based on the boundary range of the window length, the original distance diagnostic map is processed by adding a distance window to obtain the distance diagnostic map.

[0077] More specifically, the method of adding a distance window to the original distance diagnostic map can be expressed as follows:

[0078]

[0079] Where D(i) is the processed distance diagnostic map, D0 is the original distance diagnostic map, s is the window length, and j is the window number.

[0080] In the above embodiments, by performing Fourier transform and frequency coordinate to cable distance coordinate on the transformed time-domain signal to obtain the original distance diagnostic map, and then by adding a distance window to the original distance diagnostic map, the sensitivity of local defect identification can be increased, thereby improving the accuracy of fault identification results for the low-voltage cable to be identified.

[0081] In one exemplary embodiment, such as Figure 3 As shown, the low-voltage cable fault identification model is trained in the following way:

[0082] Step S310: Perform frequency domain reflection method-based testing on sample low-voltage cables with known fault types to obtain the corresponding frequency domain waveforms;

[0083] Step S311: Perform Fourier transform on the frequency domain waveform of the low-voltage cable with known fault type to obtain the distance diagnosis spectrum corresponding to the low-voltage cable with known fault type.

[0084] Step S312: Input the distance diagnostic map corresponding to the low-voltage cable with known fault types into the low-voltage cable fault identification model to be trained to obtain the predicted fault type.

[0085] Step S313: Based on the difference information between the predicted fault type and the known fault type, train the low-voltage cable fault identification model to be trained to obtain the trained low-voltage cable fault identification model.

[0086] In practice, the testing method is the same as that for the low-voltage cables to be identified; the defect samples of the low-voltage cables are tested using the frequency domain reflection method. For example... Figure 4 As shown, before the test, the cable core of the low-voltage cable defect sample at the beginning (test end) is connected to the test line of the frequency modulation signal source, the copper shield of the sample is grounded, and the end of the sample is open-circuited; then the terminal (PC) controls the output frequency (0.15-200MHz) of the frequency modulation signal source, and applies a frequency modulation voltage to the low-voltage cable defect sample through the modulation signal source to perform the test based on the frequency domain reflection method, and the reflected signal is received by the vector network analyzer.

[0087] A modulated signal V is transmitted to the head end of a low-voltage cable defect sample using an FM signal source. i (f) The reflected signal V′ is measured by a vector network analyzer. r (f), based on the transmitted modulation signal V i (f) and reflected signal V′ r (f) Obtain the reflection coefficient f represents the frequency of the test signal, and the obtained reflection coefficient Г′(f) is used as the frequency domain waveform spectrum of the low-voltage cable defect sample.

[0088] Furthermore, using the same method as the Fourier transform processing for the low-voltage cable to be identified, a Fourier transform is performed on the frequency domain waveform of the low-voltage cable with known fault types to obtain the distance diagnostic map corresponding to the low-voltage cable with known fault types. The distance diagnostic map corresponding to the low-voltage cable is used as training samples, and the known fault types are used as fault type labels to train the low-voltage cable fault identification model. More specifically, the output distance diagnostic map corresponding to the low-voltage cable is input into the low-voltage cable fault identification model to be trained, and the predicted fault type is output. Based on the difference information between the predicted fault type and the known fault type, the low-voltage cable fault identification model to be trained is trained to obtain the trained low-voltage cable fault identification model.

[0089] More specifically, the structure of the low-voltage cable fault identification model includes a cascaded input layer, a first convolutional layer, a first pooling layer, a second convolutional layer, a second pooling layer, a third convolutional layer, and a third pooling layer. The outputs of the third convolutional layer and the third pooling layer are both connected to a fourth convolutional layer, which is connected to a Softmax layer. The processing of the distance diagnostic map by the low-voltage cable fault identification model includes:

[0090] (1) Input the distance diagnosis map into the low-voltage cable fault identification model, and perform a 4*4 convolution operation using the first convolutional layer of the low-voltage cable fault identification model to obtain the first feature map.

[0091] (2) Use the first pooling layer of the low-voltage cable fault identification model to perform 2*2 edge patching and pooling operations to obtain the second feature map;

[0092] (3) Perform a 4*4 convolution operation on the second convolutional layer of the low-voltage cable fault identification model to obtain the third feature map;

[0093] (4) A 2*2 pooling operation is performed using the second pooling layer of the low-voltage cable fault identification model to obtain the fourth feature map;

[0094] (5) Perform a 4*4 convolution operation on the third convolutional layer of the low-voltage cable fault identification model to obtain the fifth feature map;

[0095] (6) Use the third pooling layer of the low-voltage cable fault identification model to perform 2*2 edge patching and pooling operations to obtain a one-dimensional sixth feature map.

[0096] (7) The fifth feature map and the one-dimensional sixth feature map are spliced ​​together using the fourth convolutional layer of the low-voltage cable fault identification model, and then convolutional operation is performed to obtain the final feature map.

[0097] (8) The Softmax layer of the low-voltage cable fault identification model is used to classify the final feature map to obtain the fault type of the low-voltage cable.

[0098] Specifically, the low-voltage cable fault identification model constructed in this embodiment includes a nine-layer structure. The first layer is the input layer, used to input distance diagnostic map samples. The input samples are grayscale images with a size of 32*32 two-dimensional data. The second layer is a convolutional layer, which generates 100 feature maps of size 29*29 after one convolution. The third layer is a pooling operation. Because a 2*2 pooling kernel is selected, the dimension of the previous convolutional layer cannot be divided evenly, so edge padding is performed, and 100 feature maps of size 15*15 are obtained after pooling. The fourth layer is a convolution operation, which generates 50 feature maps of size 12*12. The fifth layer pooling generates 50 feature maps of size 6*6, and the sixth layer convolution generates 30 feature maps of size 3*3. The seventh layer generates a one-dimensional vector through edge padding and convolution operations. The neural unit of the eighth layer consists of two parts: one part is obtained through the 1*1 convolution kernel of the seventh layer, and the other part is obtained through convolution of the feature map obtained from the sixth layer. The ninth layer is the Softmax layer, with the number of output units corresponding to the number of fault types. Instead of fully connected layers, fully convolutional layers were used in the network construction. To obtain more comprehensive features, convolutional operations were used to combine the common features of the sixth and seventh layers, which were then input into the Softmax function for classification. The aim was to improve the accuracy of fault identification.

[0099] In this embodiment, when constructing a low-voltage cable fault identification model, a fully convolutional layer is used instead of a fully connected layer, and the features of the two layers preceding the fully convolutional layer are combined, which makes the extracted fault features more sufficient and comprehensive, thereby improving the accuracy of fault identification.

[0100] In an exemplary embodiment, step S310 above performs a frequency domain reflection-based test on a sample low-voltage cable with a known fault type to obtain the corresponding frequency domain waveform, including:

[0101] Step S3101: On phase A of the exposed insulation portion of a normal low-voltage cable, a rectangular insulation layer of a set area is stripped to expose the cable core, forming a first defect. At the first defect, a test based on the frequency domain reflection method is performed between the cable core and the copper shielding layer using resistors of different resistance values ​​to obtain the frequency domain waveform corresponding to phase A.

[0102] Step S3102: Create a longitudinal knife mark defect of a set length and depth on phase B at a first distance from the cable head end of the exposed insulation part of the normal low voltage cable to form a second defect. Perform a test based on the frequency domain reflection method on the cable core and copper shielding layer at the second defect to obtain the frequency domain waveform corresponding to phase B.

[0103] Step S3103: A V-shaped defect is made along the radial direction of the cable on phase C at the second distance from the cable head end of the exposed insulation part of the normal low-voltage cable to form a third defect. At the third defect, a test based on the frequency domain reflection method is performed between the cable core and the copper shielding layer with resistors of different resistance values ​​to obtain the frequency domain waveform corresponding to C.

[0104] In this specific implementation, the cable model used in the low-voltage system of the substation is referenced. The normal low-voltage cable model is ZR-KVVP2-22 4×4 type (220 / 380V), with an insulation thickness of 0.75mm. The length L of the test cable sample is 14.1m, and the characteristic impedance of the low-voltage cable is about 50Ω.

[0105] The specific method for creating a low-voltage cable defect sample in this embodiment is as follows: A first defect (denoted as D1) is created 8.7m from the beginning (test end) of phase A of the sample (a small section of cable core is stripped from the beginning as the FDR test connection). At a distance of 8.7m from the beginning, a 10cm section of the inner and outer sheaths, steel armor, and copper shielding layer is stripped to expose the insulation. Then, a 20mm × 2mm rectangular insulation layer is stripped from the exposed insulation portion of phase A of the normal low-voltage cable to expose the cable core, forming defect D1. Figure 5 As shown. After the defect at D1 is fabricated, two transition resistors Rg with different resistance values ​​of 20Ω or 20kΩ are connected between the cable core and the copper shielding layer at D1. One end of the transition resistor is connected to the defect at D1, and the other end is grounded. This simulates low-resistance and high-resistance faults in the cable, respectively. FDR detection is performed on phase A to obtain the corresponding frequency domain waveform of phase A.

[0106] After the A-phase test is completed, a longitudinal knife mark defect of 5 mm long and 0.4 mm deep is made on the B-phase at a distance of 4.3 m from the cable end using the same method, forming a second defect (denoted as D2). That is, firstly, the outer sheath layer, steel armor, inner sheath layer, and copper shielding layer of 10 cm long at 4.3 m are peeled off in sequence. Then, the defect is made on the exposed insulation of the B-phase. After that, the FDR test is performed on the B-phase to obtain the frequency domain waveform corresponding to B.

[0107] After the B phase test is completed, a V-shaped defect along the radial direction of the cable is made on the C phase at a distance of 11m from the beginning of the cable using the same method, forming a third defect (denoted as D3). The defect depth is about 0.4mm. Then, the C phase is tested by FDR to obtain the corresponding frequency domain waveform.

[0108] refer to Figure 6 This is a schematic diagram of the frequency domain waveforms of a low-voltage cable with different resistance values ​​connected to a defect D1 in phase A, as illustrated in one embodiment. Before any resistor is connected to D1 in phase A, the amplitude of adjacent peaks in the frequency domain of the cable body decreases from the beginning to the end. When different resistance values ​​are connected to D1, the peak amplitude at D1 first increases and then decreases; that is, the peak amplitude at the defect is higher than the peaks on both sides, and the increase in peak amplitude at the defect is greater when lower resistance values ​​are connected. For example, the amplitude at D1 is -82dB when a 20kΩ resistor is connected, while it is -72dB when a 20Ω resistor is connected, which is higher than when a 20kΩ resistor is connected. Figure 6 It can be seen that the defect will cause the amplitude of the frequency domain peak at the defect location to increase and approach or exceed the peaks on both sides.

[0109] refer to Figure 7 This is a schematic diagram of the frequency domain waveforms before and after the creation of knife-mark defects D2 and D3 in phases B and C, as illustrated in an embodiment. Before the creation of knife-mark defects D2 and D3 in phases B and C, the amplitude of adjacent peaks in the frequency domain of the cable body decreased from the beginning to the end. After the creation of knife-mark defects D2 and D3, the peak amplitude at the defect is equal to or higher than the peak amplitudes on both sides of the defect (D2). After the creation of knife-mark defects D2 and D3, a positive wave appears at the defect, and the peak amplitude is slightly larger than that without knife-mark defects. Since the knife-mark defect does not cut into the cable core, it is actually a high-resistance defect. Therefore, when a high-resistance fault exists in the cable, it will manifest as an increased peak amplitude in the frequency domain, approaching or exceeding the peaks on both sides. When a low-resistance fault exists in the cable, it will also manifest as an increased amplitude in the frequency domain, but a negative wave appears in the time domain.

[0110] The testing equipment used in this embodiment is a vector network analyzer (VNA), which is used as a modulation signal source to apply a frequency-modulated voltage to the low-voltage cable under test. The voltage amplitude of the frequency-modulated signal source is generally set to 0-5V, excluding the terminal value of 0, and preferably the output voltage amplitude of the signal source is set to 5V. By applying a relatively low voltage amplitude to both ends of the low-voltage cable under test, no damage to the cable insulation will be caused, and multiple high-frequency measurements of the low-voltage cable can be achieved. The testing equipment has advantages such as small capacity and small size, and the test time required in high-frequency measurement mode is short. Therefore, this solution is applicable to engineering sites. Due to the influence of the equipment, the lower limit of the output frequency of the signal source is generally a fixed value, while the upper limit of the output frequency is related to the length of the low-voltage cable under test. The longer the low-voltage cable under test, the smaller the upper limit of the output frequency. The output frequency range set in this application is 0.15MHz-200MHz. Since the peak narrows at the defect, this setting can provide test sensitivity. The number of measurement frequency points directly affects the accuracy of FDR measurement. If the number of measurement frequency points is too small, the positioning accuracy will be low and frequency aliasing may occur, resulting in misjudgment. If the number of measurement frequency points is too large, the data processing will be complicated and the calculation time will be increased. In this application, the range of the number of measurement frequency points is 2000-4000, and preferably 3000.

[0111] In an exemplary embodiment, before inputting the distance diagnostic map corresponding to the low-voltage cable with known fault type into the low-voltage cable fault identification model to be trained in step S312 to obtain the predicted fault type, the method further includes: performing deformation processing on the distance diagnostic map of the low-voltage cable with known fault type to obtain a data-enhanced distance diagnostic map; the deformation processing includes at least one of cropping, compression and rotation.

[0112] Step S312 above also includes: inputting the data-enhanced distance diagnostic map into the low-voltage cable fault identification model to be trained to obtain the predicted fault type.

[0113] In this embodiment, by processing the distance diagnosis map of low-voltage cables with known fault types through cropping, compression, and rotation, the types of distance diagnosis maps are enriched, which can improve the versatility and recognition accuracy of the trained low-voltage cable fault identification model.

[0114] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0115] Based on the same inventive concept, this application also provides a low-voltage cable fault identification device for implementing the low-voltage cable fault identification method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more embodiments of the low-voltage cable fault identification device provided below can be found in the limitations of the low-voltage cable fault identification method described above, and will not be repeated here.

[0116] In one embodiment, such as Figure 8 As shown, a low-voltage cable fault identification device is provided, comprising: an acquisition module 810, a transformation module 820, and an identification module 830, wherein:

[0117] The acquisition module 810 is used to acquire the frequency domain waveform corresponding to the low-voltage cable to be identified; the frequency domain waveform is obtained by testing the low-voltage cable to be identified based on the frequency domain reflection method;

[0118] The transformation module 820 is used to perform Fourier transform processing on the frequency domain waveform to obtain the distance diagnostic spectrum corresponding to the low-voltage cable to be identified.

[0119] The identification module 830 is used to input the distance diagnostic map into the trained low-voltage cable fault identification model to obtain the fault type corresponding to the low-voltage cable to be identified.

[0120] In one embodiment, the transformation module 820 is specifically used to transform the reflection coefficient in the spectrum corresponding to the frequency domain waveform into a time domain signal that varies with time; perform a fast Fourier transform or a discrete Fourier transform on the real or imaginary part of the waveform spectrum of the transformed time domain signal to obtain a reflection coefficient spectrum that varies with the fundamental frequency; convert the frequency coordinates in the reflection coefficient spectrum into cable distance coordinates to obtain a distance diagnostic spectrum that varies with distance, which is used as the distance diagnostic spectrum corresponding to the low-voltage cable to be identified.

[0121] In one embodiment, the transformation module 820 is further configured to convert the frequency coordinates in the reflection coefficient spectrum into cable distance coordinates to obtain an original distance diagnostic spectrum showing the reflection coefficient as a function of distance; and to perform distance windowing processing on the original distance diagnostic spectrum to obtain a distance diagnostic spectrum showing the reflection coefficient as a function of distance.

[0122] In one embodiment, the above-mentioned apparatus further includes a training module, used to perform frequency domain reflection-based testing on sample low-voltage cables with known fault types to obtain the corresponding frequency domain waveforms; perform Fourier transform on the frequency domain waveforms of the low-voltage cables with known fault types to obtain the distance diagnostic map corresponding to the low-voltage cables with known fault types; input the distance diagnostic map corresponding to the low-voltage cables with known fault types into the low-voltage cable fault identification model to be trained to obtain the predicted fault type; and train the low-voltage cable fault identification model to be trained based on the difference information between the predicted fault type and the known fault type to obtain the trained low-voltage cable fault identification model.

[0123] In one embodiment, the device further includes a testing module for stripping a rectangular insulation layer of a predetermined area from phase A of the exposed insulation portion of a normal low-voltage cable to expose the cable core, forming a first defect. At the first defect, tests are performed using resistors of different resistance values ​​between the cable core and the copper shielding layer, based on the frequency domain reflection method, to obtain the frequency domain waveform corresponding to phase A. A longitudinal knife-mark defect of predetermined length and depth is created on phase B of the exposed insulation portion of the normal low-voltage cable at a first distance from the cable's beginning, forming a second defect. At the second defect, tests are performed using resistors of different resistance values ​​between the cable core and the copper shielding layer, based on the frequency domain reflection method, to obtain the frequency domain waveform corresponding to phase B. A V-shaped defect along the cable's radial direction is created on phase C of the exposed insulation portion of the normal low-voltage cable at a second distance from the cable's beginning, forming a third defect. At the third defect, tests are performed using resistors of different resistance values ​​between the cable core and the copper shielding layer, based on the frequency domain reflection method, to obtain the frequency domain waveform corresponding to phase C.

[0124] In one embodiment, the training module is further configured to deform the distance diagnostic map of a low-voltage cable with known fault types to obtain a data-enhanced distance diagnostic map; the deformation process includes at least one of cropping, compression, and rotation; the data-enhanced distance diagnostic map is input into the low-voltage cable fault identification model to be trained to obtain the predicted fault type.

[0125] Each module in the aforementioned low-voltage cable fault identification device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0126] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 9 As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a low-voltage cable fault identification method. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0127] Those skilled in the art will understand that Figure 9 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0128] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0129] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0130] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0131] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data shall comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0132] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0133] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0134] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A low voltage cable fault identification method characterized by, The method includes: Obtain the frequency domain waveform corresponding to the low-voltage cable to be identified; the frequency domain waveform is obtained by testing the low-voltage cable to be identified based on the frequency domain reflection method; The frequency domain waveform is subjected to Fourier transform processing to obtain the distance diagnostic spectrum corresponding to the low-voltage cable to be identified; The distance diagnostic map is input into the trained low-voltage cable fault identification model to obtain the fault type corresponding to the low-voltage cable to be identified. The low-voltage cable fault identification model includes a cascaded input layer, a first convolutional layer, a first pooling layer, a second convolutional layer, a second pooling layer, a third convolutional layer, and a third pooling layer. The outputs of the third convolutional layer and the third pooling layer are both connected to a fourth convolutional layer, which is connected to a Softmax layer. The processing procedure of the low-voltage cable fault identification model for the distance diagnostic map includes: The distance diagnostic map is input into the low-voltage cable fault identification model, and a 4*4 convolution operation is performed on the first convolutional layer of the low-voltage cable fault identification model to obtain the first feature map. The second feature map is obtained by performing a 2*2 edge patching operation and pooling operation on the first pooling layer of the low-voltage cable fault identification model. The second convolutional layer of the low-voltage cable fault identification model is used to perform a 4*4 convolution operation to obtain the third feature map; A 2*2 pooling operation is performed using the second pooling layer of the low-voltage cable fault identification model to obtain the fourth feature map; The fifth feature map is obtained by performing a 4*4 convolution operation on the third convolutional layer of the low-voltage cable fault identification model. Using the third pooling layer of the low-voltage cable fault identification model, a 2*2 edge-padding operation and pooling operation are performed to obtain a one-dimensional sixth feature map. The fifth feature map and the one-dimensional sixth feature map are concatenated using the fourth convolutional layer of the low-voltage cable fault identification model, and then convolutional to obtain the final feature map. The Softmax layer of the low-voltage cable fault identification model is used to classify the final feature map to obtain the fault type of the low-voltage cable to be identified.

2. The method of claim 1, wherein, The step of performing Fourier transform processing on the frequency domain waveform to obtain the distance diagnostic map corresponding to the low-voltage cable to be identified includes: The reflection coefficient in the spectrum corresponding to the frequency domain waveform is transformed into a time domain signal that varies with time. Perform a Fast Fourier Transform or Discrete Fourier Transform on the real or imaginary part of the waveform spectrum of the transformed time-domain signal to obtain the reflection coefficient spectrum that varies with the fundamental frequency. The frequency coordinates in the reflection coefficient spectrum are converted into cable distance coordinates to obtain a distance diagnostic spectrum of reflection coefficient as a function of distance, which is used as the distance diagnostic spectrum of the low-voltage cable to be identified.

3. The method of claim 2, wherein, The step of converting the frequency coordinates in the reflection coefficient spectrum into cable distance coordinates to obtain a distance diagnostic spectrum of reflection coefficient as a function of distance includes: The frequency coordinates in the reflection coefficient spectrum are converted into cable distance coordinates to obtain the original distance diagnostic spectrum of reflection coefficient as a function of distance; The original distance diagnostic map is processed by adding a distance window to obtain the distance diagnostic map in which the reflectance coefficient changes with distance.

4. The method according to claim 1, characterized in that, The low-voltage cable fault identification model is trained in the following manner: Test the low-voltage cables with known fault types using the frequency domain reflection method to obtain the corresponding frequency domain waveforms; Perform a Fourier transform on the frequency domain waveform of the low-voltage cable with the known fault type to obtain the distance diagnostic spectrum corresponding to the low-voltage cable with the known fault type; Input the distance diagnostic map corresponding to the low-voltage cable with the known fault type into the low-voltage cable fault identification model to be trained to obtain the predicted fault type; Based on the difference information between the predicted fault type and the known fault type, the low-voltage cable fault identification model to be trained is trained to obtain the trained low-voltage cable fault identification model.

5. The method according to claim 4, characterized in that, The step of testing sample low-voltage cables with known fault types using the frequency domain reflection method to obtain the corresponding frequency domain waveforms includes: A rectangular insulation layer of a set area is stripped from phase A of the exposed insulation portion of a normal low-voltage cable to expose the cable core, forming a first defect. At the first defect, a test based on the frequency domain reflection method is performed between the cable core and the copper shielding layer with resistors of different resistance values ​​to obtain the frequency domain waveform corresponding to phase A. A longitudinal knife mark defect of a set length and depth is made on phase B at the first distance from the cable head of the exposed insulation part of a normal low-voltage cable to form a second defect. At the second defect, a test based on the frequency domain reflection method is performed between the cable core and the copper shielding layer with resistors of different resistance values ​​to obtain the frequency domain waveform corresponding to phase B. A V-shaped defect is made along the radial direction of the cable on phase C at the second distance from the cable head end of the exposed insulation part of the normal low-voltage cable, forming a third defect. At the third defect, a test based on the frequency domain reflection method is performed between the cable core and the copper shielding layer with resistors of different resistance values ​​to obtain the frequency domain waveform corresponding to C.

6. The method according to claim 4, characterized in that, Before inputting the distance diagnostic map corresponding to the low-voltage cable with the known fault type into the low-voltage cable fault identification model to be trained to obtain the predicted fault type, the method further includes: The distance diagnostic map of the low-voltage cable with the known fault type is deformed to obtain a data-enhanced distance diagnostic map; the deformation process includes at least one of trimming, compression and rotation; The step of inputting the distance diagnostic map corresponding to the low-voltage cable with the known fault type into the low-voltage cable fault identification model to be trained to obtain the predicted fault type includes: The augmented distance diagnostic map is input into the low-voltage cable fault identification model to be trained to obtain the predicted fault type.

7. A low-voltage cable fault identification device, characterized in that, The device includes: The acquisition module is used to acquire the frequency domain waveform corresponding to the low-voltage cable to be identified; the frequency domain waveform is obtained by testing the low-voltage cable to be identified based on the frequency domain reflection method; The transformation module is used to perform Fourier transform processing on the frequency domain waveform to obtain the distance diagnostic spectrum corresponding to the low-voltage cable to be identified; The identification module is used to input the distance diagnostic map into the trained low-voltage cable fault identification model to obtain the fault type corresponding to the low-voltage cable to be identified; wherein, the low-voltage cable fault identification model includes an input layer, a first convolutional layer, a first pooling layer, a second convolutional layer, a second pooling layer, a third convolutional layer, and a third pooling layer in sequence, the outputs of the third convolutional layer and the third pooling layer are both connected to a fourth convolutional layer, and the fourth convolutional layer is connected to a Softmax layer; The identification module is specifically used to input the distance diagnostic map into the low-voltage cable fault identification model, perform a 4*4 convolution operation on the first convolutional layer of the low-voltage cable fault identification model to obtain a first feature map; perform a 2*2 edge-padding and pooling operation on the first pooling layer of the low-voltage cable fault identification model to obtain a second feature map; perform a 4*4 convolution operation on the second convolutional layer of the low-voltage cable fault identification model to obtain a third feature map; and perform a 2*2 pooling operation on the second pooling layer of the low-voltage cable fault identification model to obtain a fourth feature map. The fifth feature map is obtained by performing a 4x4 convolution operation on the third convolutional layer of the low-voltage cable fault identification model; a one-dimensional sixth feature map is obtained by performing a 2x2 edge-padding and pooling operation on the third pooling layer of the low-voltage cable fault identification model; the fifth feature map and the one-dimensional sixth feature map are concatenated by the fourth convolutional layer of the low-voltage cable fault identification model and then convolved to obtain the final feature map; the final feature map is classified by the Softmax layer of the low-voltage cable fault identification model to obtain the fault type of the low-voltage cable to be identified.

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

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

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

Images