A detection method and system for current data of AC fault series arc

Abstract

The present invention discloses a method and system for detecting current data of an AC fault series arc. Through time-frequency domain analysis of the current data sequence, the relative elevation ratio index is applied to evaluate its characteristic performance under the conditions of combined electrical apparatus disturbance with non-resistive shielding load, single and combined electrical apparatus disturbance with resistive shielding load, and key parameters such as the form of arc characteristics, decomposition scale, and characteristic frequency band are selected and determined. The optimal characteristics are used to comprehensively extract the fault information in the current data as the basis for identifying the fault arc. The method of the present invention can make the amplitude of the characteristic quantity increase significantly after the fault arc occurs, thereby quickly and accurately identifying the fault arc under the interference of complex and diverse shielding loads, and improving the safety and stable operation ability of the entire low-voltage AC circuit system.

Description

Technical Field

[0001] The present invention relates to the technical field of low-voltage AC fault arc detection, and specifically to a method and system for detecting current data of AC fault series arcs. Background Art

[0002] The problem of fault arc detection has received extensive attention from scholars in the field and has become a research hotspot. In order to improve the accuracy of fault arc detection, scholars at home and abroad have studied many methods to distinguish the normal operation state of the line from whether there is a fault, and then judge the occurrence of the arc. In order to achieve the diagnosis of fault arcs, three types of methods are generally adopted at present, mainly based on the mathematical model, physical phenomenon and electrical characteristics of the arc.

[0003] The first method is a method of identifying fault arcs by combining the mathematical model of fault arcs. Due to the complex fault arc mechanism and imperfect mathematical models, this method has developed slowly, and these models need to process too many detection parameters and overly complex identification algorithms; the second type is a method of identifying by combining the physical characteristics of fault arcs. When a fault arc occurs, abnormal behaviors such as heat increase, arc light change, current fluctuation, electromagnetic interference, and radiation caused by discharge will occur, and these phenomena can be used to judge the fault arc. Although specific physical behaviors can detect fault arcs, the diversity of load properties and the fixed installation position of sensors limit the application of these characteristics due to the uncertainty of the fault point position. For this reason, the convenience of circuit current measurement makes it an ideal parameter for fault arc diagnosis; the third type is the identification method based on the change of loop electrical signal characteristics of fault arcs. Due to the ineffectiveness of voltage sampling in a large range, there are disadvantages in the application of voltage characteristic changes in practice. The fault arc current is independent of the position where it is generated, so it has become the most common method to detect fault arcs by measuring the current. Current signals are usually processed in the time domain, frequency domain or a combination of the two characteristics.

[0004] With the progress of information technology, the latest development of artificial intelligence (AI) provides new ideas for fault arc detection. Machine learning (ML) methods have gradually become popular in fault diagnosis technologies in various fields, and some recent studies in the field of fault arc detection have achieved good results. In addition to effective pattern recognition capabilities, machine learning can also be robust and adaptable in systems with general operating conditions. The significant difference between AI-based methods and preset threshold-based methods is that the boundary separating fault arcs from normal operations is determined by the training process rather than manual processing. Most of the above detection algorithm effects are only obtained through theoretical calculations and have not been effectively experimentally verified in the real system operating environment. For the detection module that has been implemented in hardware, due to the incomplete coverage of load types involved in the research and the lack of effective and sufficient research on crosstalk interference, etc., it leads to the situation of circuit disconnection due to misjudgment under different shielded load conditions, causing unnecessary losses. Summary of the Invention

[0005] Aiming at the problems existing in the prior art, the present invention provides a detection method and system for AC fault series arcs, which can accurately detect fault arcs in current data according to the optimal feature quantity.

[0006] The present invention is realized through the following technical solutions:

[0007] A detection method for current data of AC fault series arcs includes the following steps:

[0008] Step 1: Use the relative lift ratio index to determine the final optimal feature quantity of current data under the conditions of non-resistive shielded load and resistive shielded load, and single electrical appliance and combined electrical appliance perturbation conditions;

[0009] Step 2: Perform wavelet decomposition on the final optimal feature quantity and determine the optimal decomposition layer. According to the optimal decomposition layer, determine the optimal feature frequency band, and perform fault arc detection on the current data according to the eigenvalue of the feature quantity corresponding to the optimal feature frequency band.

[0010] Preferably, the method for determining the final optimal feature quantity in Step 1 is as follows:

[0011] Determine the lift ratio of each feature quantity of current data under the conditions of non-resistive shielded load and resistive shielded load, and single electrical appliance and combined electrical appliance perturbation conditions. According to the lift ratio of each feature quantity under each perturbation condition, determine the optimal feature quantity under each perturbation condition. According to the optimal feature quantity under each perturbation condition, determine the final optimal feature quantity of current data.

[0012] Preferably, perform time-frequency domain analysis on the sequence of current data to obtain each feature quantity of the current data.

[0013] Preferably, the method for determining the optimal feature quantity is as follows:

[0014] For the improvement ratios obtained under each perturbation condition, all feature quantities are normalized based on the feature quantity with the minimum improvement ratio, and the relative improvement ratios of each feature quantity with respect to the feature quantity with the minimum improvement ratio are obtained. The feature quantity corresponding to the maximum relative improvement ratio is used as the optimal feature quantity under this perturbation condition.

[0015] Preferably, the calculation method of the improvement ratio is as follows:

[0016] K = k2 / k1

[0017] Wherein, k2 is the average value of the current feature quantity of the faulty arc; k1 is the average value of the current feature quantity during normal operation.

[0018] Preferably, the method for determining the final optimal feature quantity of the current data is as follows:

[0019] The optimal feature quantities under each perturbation condition are statistically analyzed, and the optimal feature quantity that appears is used as the final optimal feature quantity.

[0020] Preferably, when the optimal feature quantities under each perturbation condition are all different, the relative improvement ratios of the feature quantities under each perturbation condition are sorted from smallest to largest, and the feature quantities at the next level of the optimal feature quantity under each perturbation condition are statistically analyzed. The feature quantity that appears the most among the next-level feature quantities is used as the final optimal feature quantity.

[0021] Preferably, the method for determining the optimal decomposition layer in step 2 is as follows:

[0022] According to the set detection time, the maximum wavelet decomposition layer of the final optimal feature quantity is determined. The final optimal feature quantity is decomposed layer by layer to the maximum layer, and the improvement ratio of the feature quantity of each decomposition layer is determined. The decomposition layer corresponding to the highest improvement ratio is used as the optimal decomposition layer.

[0023] Preferably, the method for determining the optimal feature frequency band in step 2 is as follows:

[0024] According to the multiple feature frequency bands of the optimal decomposition layer, the improvement ratio of the feature quantity of each feature frequency band is determined. The feature frequency band corresponding to the feature quantity with the maximum improvement ratio is used as the optimal feature frequency band.

[0025] A system for detecting the current data of an AC fault series arc, including

[0026] An optimization feature quantity module, which is used to determine the final optimization feature quantity of the current data under the perturbation conditions of single electrical appliances and combined electrical appliances when there is no resistive shielding load and there is a resistive shielding load by using the relative improvement ratio index;

[0027] A diagnostic module, which is used to perform wavelet decomposition on the final optimal feature quantity output by the optimal feature quantity module, determine the optimal decomposition layer, determine the optimal feature frequency band according to the optimal decomposition layer, and detect fault arcs in the current data based on the eigenvalue of the feature quantity corresponding to the optimal feature frequency band.

[0028] Compared with the prior art, the present invention has the following beneficial technical effects:

[0029] The present invention discloses a method for detecting current data of an AC fault series arc, performs time-frequency domain analysis on the current data sequence, evaluates its characteristic performance under the conditions of combined electrical appliance disturbance with resistive shielding load and single and combined electrical appliance disturbances with resistive shielding load by applying the relative lift ratio index, selects and determines key parameters such as arc characteristic form, decomposition scale, and characteristic frequency band, and uses the optimal features to comprehensively extract the fault information in the current data as the basis for identifying fault arcs. The method of the present invention can make the amplitude of the characteristic quantity increase significantly after the occurrence of the fault arc, thereby quickly and accurately identifying the fault arc under complex and diverse shielding load interferences, and improving the safe and stable operation ability of the entire low-voltage AC circuit system. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a structural diagram of the fault arc simulation system of the present invention;

[0031] Figure 2 It is a structural diagram of the arc simulation model subsystem of the present invention;

[0032] Figure 3 It is a simulation waveform of a series fault arc under the condition of a 30Ω resistive load of the present invention;

[0033] Figure 4 It is an arc simulation model under the influence of a 30Ω + 100mH resistive-inductive load of the present invention;

[0034] Figure 5 It is an arc simulation model under the influence of a single-phase induction motor load of the present invention;

[0035] Figure 6 It is an arc simulation model under the influence of a power electronic load of the present invention;

[0036] Figure 7 It is an experimental circuit for studying the characteristics of low-voltage fault arcs of the present invention;

[0037] Figure 8 It is an AC fault arc current waveform under the condition of resistive load operation of the present invention;

[0038] Figure 9 It is an AC fault arc current waveform under the condition of resistive load operation of the present invention;

[0039] Figure 10 For the ratio of the improvement of the fault arc characteristics of the combined electrical apparatus by different characteristic quantities under the condition that no shielding load is added to the present invention;

[0040] Figure 11 For the ratio of the improvement of the fault arc characteristics of a single electrical apparatus by different characteristic quantities under the condition that a resistive shielding load is added to the present invention;

[0041] Figure 12 For the ratio of the improvement of the fault arc characteristics of the combined electrical apparatus by different characteristic quantities under the condition that a resistive shielding load is added to the present invention;

[0042] Figure 13 For the ratio of the improvement of the fault arc characteristics under different decomposition layer numbers of the present invention;

[0043] Figure 14 For the optimized fault arc characteristic detection effect under the TV load condition of the present invention;

[0044] Figure 15 For the flowchart of the detection method of the current data of the AC fault series arc of the present invention. Detailed implementation manners

[0045] The present invention will be further described in detail below with reference to the accompanying drawings. The following is an explanation rather than a limitation of the present invention.

[0046] Refer to Figure 15 , a detection method for the current data of an AC fault series arc, comprising the following steps:

[0047] Step 1, obtain the current data of the AC fault series arc through an arc physical modeling and experimental diagnosis multi-source generation method;

[0048] Specifically, use the Cassie arc model to establish a series fault arc simulation system. The series fault arc simulation system is as Figure 1 shown. The load types considered by the fault arc simulation system include resistive load, inductive load, motor load, and power electronics load. Figure 2 is the arc simulation model subsystem. The arc simulation model determines the current parameter to be 0-63 A and the line impedance parameter to be 0-100 m through the change of source-load parameters.

[0049] The simulation waveforms of four typical loads simulated by the series fault arc simulation system are as Figures 3 - 6 shown. When the system is operating normally, its waveform is a standard sine wave; after a fault arc occurs, the current waveform morphology will show different change forms. Based on this, the load matching classification of the simulation data and the experimental data can be realized, and an AC fault arc current database in a multi-source generation method can be formed. From Figure 3From the simulation waveforms of series fault arcs under resistive load conditions, it can be seen that when an arc fault occurs on the line, the current signal in the line will experience a zero-crossing rest phenomenon every time it passes through zero. Due to the inductance in the inductive load having a certain inhibitory effect on current mutation, such as Figure 4 the zero-crossing rest phenomenon of the current signal in the line at zero-crossing is no longer significant, and only a small mutation pulse appears. Such as Figure 5 The induction motor still belongs to the resistive-inductive load, so the small pulse phenomenon still exists. When an arc fault occurs, the starting current of the single-phase induction motor load is lower than that in the normal state and the time increases. Due to the influence of the IGBT delay conduction effect, such as Figure 6 the zero-crossing time of the fault arc current is more obvious.

[0050] According to Figure 7 shown in the reference standard GB / T 31143, an experimental platform was built, and the experimental waveforms of fault arc currents were obtained according to the load categories. The experimental loads mainly included 15 types of key resistive, inductive, motor and power electronic loads, as well as loads of the same type but different brands. Based on the best simulation of the experiment, the arc model parameters were obtained: the arc time constant was taken as 2.25×10 -4 s, the arc voltage constant was taken as uc = 50V, and the initial value of the arc conductance was taken as g(0) = 1.17×10 -4 s. The experimental arc data of the finite load were used to debug the fault arc model parameters until the experimental and simulation fitting degree R2 reached more than 0.95;

[0051] For Figure 1 the results of the simulation system shown were verified for effectiveness. The experimental results of AC arcs for resistive loads and motor loads are as Figures 8 - 9 shown. As can be seen from the figure, the experimental morphology of the AC arc corresponds to the simulation results, confirming the effectiveness of the simulation model; at the same time, both the experimental and simulation results show that there are specific differences in the current characteristics and morphology of arcs under different loads, which can be used to match and form a fault arc current database.

[0052] Using the above experimental and simulation platform, fault arc data under various load influence conditions were obtained. According to the simulation model, an arc database was formed for four types of loads. The obtained experimental fault arc data and each type of simulation signal were used to calculate the correlation degree by signal correlation analysis, and the highest correlation was matched to the arc database in the corresponding load type.

[0053] Step 2: Use the relative improvement ratio index to determine the final optimized characteristic quantity of the current data under the conditions of no resistive shielding load and with resistive shielding load, and under the disturbance conditions of single electrical appliances and combined electrical appliances.

[0054] Perform time-frequency domain analysis on the current data sequence to obtain the characteristic quantities of the current data, calculate the improvement ratios of each characteristic quantity under the disturbance conditions of a single electrical appliance and a combined electrical appliance when there is no resistive shielding load; and the improvement ratios of each characteristic quantity under the disturbance conditions of a single electrical appliance and a combined electrical appliance when there is a resistive shielding load;

[0055] For the improvement ratios obtained under each disturbance condition, normalize all the characteristic quantities based on the characteristic quantity with the smallest improvement ratio to obtain the relative improvement ratios of each characteristic quantity with respect to the characteristic quantity with the smallest improvement ratio. Take the characteristic quantity corresponding to the largest relative improvement ratio as the optimal characteristic quantity under this disturbance condition. Statistically analyze the optimal characteristic quantities under each disturbance condition, and take the optimal characteristic quantity that appears as the final optimal characteristic quantity.

[0056] For example, there are a total of four working conditions: the disturbance condition of a single electrical appliance when there is no resistive shielding load, the disturbance condition of a combined electrical appliance when there is no resistive shielding load, the disturbance condition of a single electrical appliance when there is a resistive shielding load, and the disturbance condition of a combined electrical appliance when there is a resistive shielding load. If the optimal characteristic quantity for 3 of the disturbance conditions is rbio3.1, then rbio3.1 is the final optimal characteristic quantity.

[0057] When the optimal characteristic quantities under each disturbance condition are all different, sort the relative improvement ratios of the characteristic quantities under each disturbance condition from smallest to largest. Statistically analyze the characteristic quantities at the next level of the optimal characteristic quantity under each disturbance condition, that is, the characteristic quantities adjacent to the optimal characteristic quantity. Take the characteristic quantity that appears the most among the next-level characteristic quantities as the final optimal characteristic quantity.

[0058] Statistically analyze the optimal characteristic quantities under each disturbance condition. When the number of two optimal characteristic quantities is the same, calculate the difference in the improvement ratios of each optimal characteristic quantity A and characteristic quantity B under the same disturbance condition. Characteristic quantity B is the characteristic quantity with the same attribute as another optimal characteristic quantity. Take the optimal characteristic quantity with the smallest difference in improvement ratios as the final optimal characteristic quantity.

[0059] The extraction of the final optimal characteristic quantity is a key factor affecting the accurate diagnosis of current data. The information obtained from the final optimal characteristic quantity should be sufficient to express the current information of the normal operation, fault conditions, and transient processes of the power system. The present invention performs time-frequency domain analysis on the above current data, and then compares and selects the optimal characteristic quantity to ensure the comprehensive extraction of characteristic information in the current data for accurate fault detection.

[0060] The calculation formula for the improvement ratio is as follows:

[0061] K = k2 / k1

[0062] In Equation 1:

[0063] k2——The average value of the current characteristic quantity of the fault arc;

[0064] k1——Average value of current characteristic quantity during normal operation.

[0065] After calculating the multi - feature promotion ratio, all features are normalized based on the feature with the smallest promotion ratio to obtain the relative promotion ratio of each feature quantity to the smallest feature quantity. If the value of the relative promotion ratio K is larger, it indicates that the distinction of the current characteristic values before and after the occurrence of the fault arc is more obvious, and it is easier to locate the occurrence time of the fault arc. Correspondingly, we hope that the relative promotion ratio of the current characteristics during the transient process of the system is as small as possible to facilitate accurately identifying whether the fault arc actually occurs.

[0066] Taking five feature quantities, namely Sym 4, Db 2, Coif 2, Bior 3.3, and Rbio 3.1, as examples to illustrate the method for determining the optimal feature quantity. Calculate the alternating - current fault - arc current under the interference condition of a single load such as a small sun, obtain the feature results and calculate their promotion ratios. To reduce the interference of manual operation on the current data during the simulation test, the current data of the load arc is calculated ten times repeatedly and the mean value is taken. Calculate the promotion ratios under 16 load conditions in sequence, and then normalize them according to the bior3.3 feature to obtain the relative promotion ratios as Figure 10 shown. Among them, the promotion ratio of rbior3.1 is the largest; the effect of sym4 is the second, with fluctuations; the promotion ratios of db2 and coif2 are similar; the promotion ratio of bior3.3 is the smallest. Considering the influence of the resistive shielding load, first consider the relative promotion ratio under the interference condition of a single load. Among them, the promotion ratio of rbior3.1 is the largest; the effect of db2 is the second, with fluctuations; the promotion ratios of sym4 and coif2 are similar; the promotion ratio of bior3.3 is the smallest. Secondly, consider the relative promotion ratio under the interference condition of combined loads. The promotion ratio of rbior3.1 is the largest; the effect of db2 is the second; the promotion ratios of sym4 and coif2 are similar; the promotion ratio of bior3.3 is the smallest. Then select the rbio3.1 wavelet as the final optimal feature quantity.

[0067] Step 3: Perform wavelet decomposition on the final optimal feature quantity and determine the optimal decomposition layer. Determine the optimal feature frequency band according to the optimal decomposition layer, and perform fault - arc detection on the current data based on the characteristic values of the feature quantity corresponding to the optimal feature frequency band.

[0068] The method for determining the optimal decomposition layer is as follows:

[0069] Determine the maximum wavelet decomposition layer of the final optimal feature quantity according to the set detection time. Decompose the final optimal feature quantity layer by layer to the maximum layer, determine the promotion ratio of the feature quantity of each decomposition layer, and take the decomposition layer corresponding to the highest promotion ratio as the optimal decomposition layer.

[0070] The method for determining the optimal feature frequency band is as follows:

[0071] According to multiple characteristic frequency bands of the optimal decomposition level, determine the improvement ratio of the characteristic quantity of each characteristic frequency band, and use the characteristic frequency band corresponding to the characteristic quantity with the largest improvement ratio as the optimal characteristic frequency band.

[0072] When performing wavelet decomposition, minimize the detection time (i.e., reduce the wavelet decomposition level) as much as possible and have a good characteristic improvement ratio. After selecting an appropriate decomposition scale, by observing the decomposed frequency bands respectively, determine the optimal characteristic frequency band. Determine key parameters such as the decomposition scale and the characteristic frequency band, and use the optimal characteristics to comprehensively extract the fault information in the current data as the basis for discriminating the fault arc.

[0073] Different wavelet decomposition levels will affect the range of the analysis frequency band. If the wavelet decomposition level is too low, it will lead to an overly large wavelet transform analysis frequency band range, which is likely to introduce interference such as environmental noise and load noise, and is not conducive to feature construction; if the wavelet decomposition level is too high, it will cause an increase in the code operation amount and prolong the detection time. Therefore, it is necessary to minimize the wavelet decomposition level to the greatest extent within an acceptable characteristic improvement ratio range.

[0074] In this embodiment, the improvement effect of the minimum decomposition level 2 is set to 1. As Figure 13 shown, as the wavelet decomposition level increases, the characteristic improvement ratio gradually increases. Taking the characteristic improvement ratio of 2-layer decomposition as 1, the change rate of the improvement ratio of 4-layer decomposition increases significantly, and the change rate of the improvement ratio of 7-layer decomposition begins to slow down. After comprehensive comparison, 6-layer wavelet decomposition is adopted.

[0075] The present invention also provides a system for a detection method of current data of an AC fault series arc, including an optimal characteristic quantity module and a diagnosis module.

[0076] The optimal characteristic quantity module is used to determine the final optimal characteristic quantity of the current data under the conditions of non-resistive shielding load and resistive shielding load, and single electrical appliance and combined electrical appliance disturbance conditions by using the relative improvement ratio index.

[0077] The diagnosis module is used to perform wavelet decomposition on the final optimal characteristic quantity output by the optimal characteristic quantity module and determine the optimal decomposition layer, determine the optimal characteristic frequency band according to the optimal decomposition layer, and perform fault arc detection on the current data according to the characteristic value of the characteristic quantity corresponding to the optimal characteristic frequency band.

[0078] According to the optimized wavelet basis, decomposition level and frequency band as described above, the obtained fault arc wavelet characteristics are as Figure 14As shown, it can be seen that the amplitudes of the characteristic quantities before and after the occurrence of the faulty arc have a relatively obvious increase. For the transient process of the system, the characteristics can also be clearly identified and distinguished, indicating that the characteristic quantities constructed by using wavelet transform have good applicability. The method in the present invention can make the amplitude of the characteristic quantity increase significantly after the occurrence of the faulty arc, thereby quickly and accurately identifying the faulty arc under complex and diverse shielded load interferences and improving the safe and stable operation ability of the entire low-voltage AC circuit system.

[0079] The above content is only to illustrate the technical idea of the present invention and cannot be used to limit the protection scope of the present invention. Any modification made on the basis of the technical solution according to the technical idea proposed by the present invention falls within the protection scope of the claims of the present invention.

Claims

1. A method for detecting current data of an AC fault series arc, characterized in that, Including the following steps: Step 1: Use the relative lift ratio index to determine the final optimal characteristic quantities under the conditions of non-resistive shielding load and resistive shielding load, for single electrical appliances and combined electrical appliances under disturbance conditions; The method for determining the optimal characteristic quantities is as follows: For the lift ratios obtained under each disturbance condition, normalize all the characteristic quantities based on the characteristic quantity with the smallest lift ratio, obtain the relative lift ratios of each characteristic quantity with respect to the characteristic quantity with the smallest lift ratio, and take the characteristic quantity corresponding to the largest relative lift ratio as the optimal characteristic quantity under this disturbance condition; The method for determining the final optimal characteristic quantities is as follows: Determine the lift ratios of each characteristic quantity of the current data under the conditions of single electrical appliances and combined electrical appliances under non-resistive shielding load and resistive shielding load, determine the optimal characteristic quantity under each disturbance condition according to the lift ratios of each characteristic quantity under each disturbance condition, and determine the final optimal characteristic quantity of the current data according to the optimal characteristic quantities under each disturbance condition; The method for determining the final optimal characteristic quantity of the current data is as follows: Statistically analyze the optimal characteristic quantities under each disturbance condition, and take the occurring optimal characteristic quantity as the final optimal characteristic quantity; The calculation method of the lift ratio is as follows: Among them, k 2 is the average value of the current characteristic quantity of the faulty arc; k 1 is the average value of the current characteristic quantity during normal operation; When the optimal characteristic quantities under each disturbance condition are all different, sort the relative lift ratios of the characteristic quantities under each disturbance condition from smallest to largest, statistically analyze the characteristic quantities at the next level of the optimal characteristic quantity under each disturbance condition, and take the characteristic quantity that appears the most among the next-level characteristic quantities as the final optimal characteristic quantity; Step 2: Perform wavelet decomposition on the final optimal characteristic quantity and determine the optimal decomposition layer, determine the optimal characteristic frequency band according to the optimal decomposition layer, and perform fault arc detection on the current data according to the eigenvalue of the characteristic quantity corresponding to the optimal characteristic frequency band.

2. The detection method of the current data of an AC fault series arc according to claim 1, characterized in that Perform time-frequency domain analysis on the sequence of the current data to obtain each characteristic quantity of the current data.

3. The detection method of current data of an AC fault series arc according to claim 1, characterized in that The method for determining the optimal decomposition layer in Step 2 is as follows: Determine the maximum wavelet decomposition layer of the final optimal characteristic quantity according to the set detection time, decompose the final optimal characteristic quantity layer by layer to the maximum layer, determine the lift ratio of the characteristic quantity of each decomposition layer, and take the decomposition layer corresponding to the highest lift ratio as the optimal decomposition layer.

4. The detection method of the current data of an AC fault series arc according to claim 3, characterized in that, The method for determining the optimal characteristic frequency band in Step 2 is as follows: Determine the lift ratio of the characteristic quantity of each characteristic frequency band according to the multiple characteristic frequency bands of the optimal decomposition layer, and take the characteristic frequency band corresponding to the characteristic quantity with the largest lift ratio as the optimal characteristic frequency band.

5. A system for detecting current data of an AC fault series arc according to any one of claims 1-4, characterized in that, Including An optimal characteristic quantity module, which is used to use the relative lift ratio index to determine the final optimal characteristic quantities of the current data under the conditions of non-resistive shielding load and resistive shielding load, for single electrical appliances and combined electrical appliances under disturbance conditions; A diagnosis module, which is used to perform wavelet decomposition on the final optimal characteristic quantity output by the optimal characteristic quantity module and determine the optimal decomposition layer, determine the optimal characteristic frequency band according to the optimal decomposition layer, and perform fault arc detection on the current data according to the eigenvalue of the characteristic quantity corresponding to the optimal characteristic frequency band.

Images