Power distribution network arc fault detection method based on power disturbance modeling

Through the method based on power disturbance modeling, the distribution network and disturbance model are constructed, the analysis model is built and the parameter fitting error judgment is carried out, which solves the problem of low accuracy in arc fault detection in the existing technology, and achieves high sensitivity to arc faults Identify and accurately distinguish.

CN118962329APending Publication Date: 2024-11-15YIBIN POWER SUPPLY COMPANY STATE GRID SICHUAN ELECTRIC POWER
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411032881.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-30
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

The existing distribution network arc fault detection methods have low accuracy, and cannot effectively distinguish arc and non-arc faults, and ignore the nonlinear characteristics of arc and the influence of environmental factors.

Method used

Based on power disturbance modeling, a distribution network model and disturbance model are constructed, non-arc and arc disturbance analysis models are built, and the mathematical model and parameter fitting errors are judged. Parameters are estimated using the lsqcurvefit function, and binary hypothesis test is performed to identify arc faults.

Benefits of technology

It improves the sensitivity and accuracy of arc fault detection, comprehensively considers the judgment of non-arc and arc faults, accurately reflects the nonlinear characteristics of arc, and improves detection accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118962329A_ABST
    Figure CN118962329A_ABST
Patent Text Reader

Abstract

The invention discloses a power distribution network arc fault detection method based on electric power disturbance modeling, and relates to the technical field of power distribution network arc fault detection, and the method comprises the following steps: constructing a distribution network model, and constructing a disturbance model based on the distribution network model; applying the disturbance model to the distribution network model, and respectively building a non-arc disturbance analysis model and an arc disturbance analysis model; and based on the non-arc disturbance analysis model and the arc disturbance analysis model, a fault detection model is used to determine a measurement signal, and power distribution network arc fault detection based on electric power disturbance modeling is completed. The problems that an existing arc fault detection method is limited in use and low in detection accuracy are solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of arc fault detection in distribution networks, and in particular to a method for detecting arc faults in distribution networks based on power disturbance modeling. Background Art

[0002] At present, the proportion of underground cables in urban distribution networks is increasing. Due to the need for power supply reliability, neutral point non-effective grounding is often used in urban distribution networks. Among the types of cable faults, more than 80% are single-phase grounding faults. When a single-phase fault occurs in a neutral point non-effective grounding system, it is easy to form an intermittent fault arc that alternates between extinguishing and reigniting, posing a serious threat to the normal operation of the power grid.

[0003] At present, one way to detect arc faults is to use wavelet decomposition to detect the zero-off period of the arc. However, there are two views on the extinction of the arc. One is that the arc can be extinguished when the high-frequency oscillating current passes through zero during the transition process, and the other is that the arc cannot be extinguished until the power frequency current passes through zero. In addition, the humidity of the environment and the arc burning characteristics of the arc also have a great influence on the "zero-off", and the use of the wavelet detection method is limited; the second is to detect arc faults based on the voltage waveform distortion rate, but ignore the influence of non-arc fault grounding impedance and arc dissipation power changes; in addition, there is also the identification of fault types by setting thresholds for the characteristics of electrical quantity changes generated during the fault period, but there is still a lack of consideration of the nonlinear characteristics of the arc. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides a distribution network arc fault detection method based on power disturbance modeling, which solves the problems of limited use and low detection accuracy of the existing arc fault detection methods.

[0005] In order to achieve the above-mentioned invention object, the technical solution adopted by the present invention is: a distribution network arc fault detection method based on power disturbance modeling, comprising the following steps:

[0006] S1: Build a distribution network model, and build a disturbance model based on the distribution network model;

[0007] S2: Apply the disturbance model to the distribution network model, and build a non-arc disturbance analysis model and an arc disturbance analysis model respectively;

[0008] S3: Based on the non-arcing disturbance analysis model and the arc disturbance analysis model, the fault detection model is used to judge the measurement signal to complete the distribution network arc fault detection based on power disturbance modeling.

[0009] Furthermore, the distribution network model in S1 is:

[0010]

[0011] Among them, v s is the station voltage, i s is the station current, R eq is the equivalent resistance between the substation and the disturbance current, L eq is the equivalent inductance between the substation and the disturbance, v f is the voltage at the disturbance point, and t is the time.

[0012] Furthermore, the disturbance model in S1 includes a non-arcing disturbance model and an arc fault model;

[0013] The non-arc disturbance model adopts a non-arc disturbance general RL model;

[0014] The arc fault model is:

[0015]

[0016] u st =u0+r0|i f

[0017] Where g is the dynamic arc conductance, τ is the time constant of the arc, G is the steady-state arc conductance, i f is the disturbance current, u st is the steady-state arc voltage, u0 is the characteristic arc voltage, and r0 is the characteristic arc resistance.

[0018] Furthermore, the non-arcing disturbance analysis model in S2 is:

[0019]

[0020] Where, R is the static non-arcing resistance, L is the static non-arcing inductance;

[0021] The arc disturbance analysis model is:

[0022]

[0023] Among them, R0 is the constant resistance of the arc.

[0024] Furthermore, in S3, the fault detection model is used to judge the measurement signal, including the following steps:

[0025] S3-1: Estimate the load impedance on the transmission line based on the station-end voltage and station-end current before the disturbance;

[0026] S3-2: Estimate the load current and estimate the disturbance current in combination with KCL;

[0027] S3-3: Based on the estimated load impedance, load current and disturbance current, the non-arcing disturbance parameters and arc fault parameters are estimated using the lsqcurvefit function respectively;

[0028] S3-4: recalculating more accurate model parameters and disturbance data based on the estimated non-arcing disturbance parameters and arc fault parameters;

[0029] S3-5: Based on more accurate model parameters and disturbance data, the curve fitting error is used to perform binary hypothesis testing on the results to complete the judgment of the measurement signal.

[0030] Furthermore, the load impedance Z in S3-1 load Estimated to be:

[0031]

[0032] Among them, R load and L load is the load impedance parameter, j is the imaginary part of the load impedance, and f is the grid frequency.

[0033] Furthermore, the load current estimation i in S3-2 load (n) is:

[0034]

[0035] Among them, v s (n) is the value of the station voltage at time nΔt, Δt is the sampling time interval, n is a positive integer, i load (n-1) is the load current at the previous moment;

[0036]

[0037] Among them, i load is the load current;

[0038] The disturbance current is estimated as:

[0039] i f (n) = i s (n)-i load (n)

[0040] Among them, i f (n) is the value of the disturbance current at time nΔt, i s (n) is the value of the station current at time nΔt;

[0041] The differential equation of the disturbance current is:

[0042]

[0043] Furthermore, in S3-3, the non-arcing disturbance parameters and arc fault parameters are estimated respectively using the lsqcurvefit function, including:

[0044] Estimate the unknown parameters under the assumption of non-arcing disturbance based on the least squares method:

[0045]

[0046] in, is the estimated value of the non-arcing disturbance parameter, n0 is the sampling start time, N is the number of sampling points, F0(·) is the estimated station voltage under the non-arcing disturbance assumption, x is the non-arcing disturbance model parameter, and data(n) is the known parameter vector in the non-arcing disturbance;

[0047] Based on the lsqcurvefit function, the non-arcing disturbance parameter estimation results are obtained

[0048] Estimate the unknown parameters under the assumption of arc fault based on the least squares method:

[0049]

[0050] in, is the estimated value of arc fault parameters, F1(·) is the estimated station terminal voltage under the assumption of arc fault;

[0051] The arc fault model is:

[0052]

[0053] Wherein, g(n) is the arc conductance at time nΔt, and g(n-1) is the arc conductance at the previous time;

[0054] The arc conductance g(n) is:

[0055]

[0056] Based on the lsqcurvefit function, the arc fault parameter estimation results are obtained

[0057] The superscript ∧ indicates an estimated value.

[0058] Further, the curve fitting error in S3-5 includes the curve fitting error under the non-arcing disturbance assumption and the curve fitting error under the arc fault assumption;

[0059] The curve fitting error e under the assumption of no arc disturbance non-arc for:

[0060]

[0061] According to the non-arcing disturbance parameter estimation results Calculate the station voltage under the assumption of no arc disturbance for:

[0062]

[0063] The curve fitting error e under the arc fault assumption condition arc for:

[0064]

[0065] According to the arc fault parameter estimation results Calculation of the station voltage under the assumption of an arc fault for:

[0066]

[0067] Furthermore, in S3-5, the curve fitting error is used to perform a binary hypothesis test on the results, and the formula is:

[0068]

[0069] Among them, H0 represents non-arcing disturbance and H1 represents arc fault.

[0070] The beneficial effects of the present invention are as follows: the present invention comprehensively considers the discrimination between non-arc faults and arc faults, and accurately reflects the nonlinear characteristics of arcs through mathematical models. Based on the idea of ​​parameter fitting, the fitting errors of arc fault models and non-arc fault models are used to determine arc faults; the present invention uses the line impedance identified by parameters to update and calculate the load impedance, and improves the fitting accuracy of the arc fault model through the iteration of load impedance, thereby improving the sensitivity of arc fault detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 The flowchart of a distribution network arc fault detection method based on power disturbance modeling is shown in FIG.

[0072] Figure 2 This is a simplified distribution network model structure diagram.

[0073] Figure 3 This is the structure diagram of the general RL model for non-arcing disturbances.

[0074] Figure 4 This is the structural diagram of the dynamic resistance model of arc fault.

[0075] Figure 5 This is the structure diagram of the distribution network non-arcing disturbance and arcing disturbance model. DETAILED DESCRIPTION

[0076] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0077] like Figure 1 As shown, a distribution network arc fault detection method based on power disturbance modeling includes the following steps:

[0078] S1: Build a distribution network model, and build a disturbance model based on the distribution network model;

[0079] S2: Apply the disturbance model to the distribution network model, and build a non-arc disturbance analysis model and an arc disturbance analysis model respectively;

[0080] S3: Based on the non-arcing disturbance analysis model and the arc disturbance analysis model, the fault detection model is used to judge the measurement signal to complete the distribution network arc fault detection based on power disturbance modeling.

[0081] In one embodiment of the present invention, in order to realize model-based cable arc fault detection, it is first necessary to build a reasonable distribution network model. Taking the urban 10kV distribution network as an example, it mainly includes substations, transmission lines and loads, where the substation model can be represented by an AC equivalent power source. Considering that the cable line is not long and the operating voltage level is not high, the influence of conductivity and susceptance can be ignored and represented by an impedance model. The load model mainly includes a constant power load model and a constant impedance load model. Considering that this patent involves short-circuit calculations, the load model is represented by equivalent impedance. Since the actual distribution network only contains the station-end voltage and current information of the substation and the large number of lines and branch loads is not conducive to fault identification, it needs to be appropriately simplified. The following can be used: Figure 2 The distribution network model shown in the figure, where R eq and L eq It can be understood as the equivalent resistance and equivalent inductance between the substation and the disturbance. All loads on the feeder are represented by Z load The resistance part is represented by R load , the inductance part is represented by L load .

[0082] The distribution network model in S1 is:

[0083]

[0084] Among them, v s is the station voltage, i s is the station current, R eq is the equivalent resistance between the substation and the disturbance current, L eq is the equivalent inductance between the substation and the disturbance, v f is the voltage at the disturbance point, and t is the time.

[0085] like Figure 2 As shown, it is necessary to establish a disturbance model at the disturbance point, which is mainly divided into arc faults and non-arc disturbances. For non-arc disturbances, they mainly include the start-up of the unit, the switching of the load, and constant impedance faults. These disturbances will cause a sudden increase in the station-end current. For the start-up of the generator set, because its dynamic process usually only lasts for a few seconds, it can be regarded as a constant impedance model when only a short time scale (ms level) is considered; for the switching of the load, although there is a nonlinear load part, because the linear load accounts for the main part in practice, the influence of the nonlinear load is weakened and ignored. In summary, in most cases, the following can be used Figure 3 The static impedance model shown is used to characterize their common characteristics.

[0086] For the arc model, in order to restore the nonlinear time-varying characteristics of the arc burning in the air, the energy balance of the arc column is taken into account, and the Kizilcay model of the arc in the air is described by the differential equation of the arc conductance. Therefore, the arc fault model can be adopted as follows Figure 4 The dynamic resistance model is shown.

[0087] The disturbance model in S1 includes non-arcing disturbance model and arc fault model;

[0088] The non-arc disturbance model adopts a non-arc disturbance general RL model;

[0089] The arc fault model is:

[0090]

[0091] u st =u0+r0|i f

[0092] Where g is the dynamic arc conductance, τ is the time constant of the arc, G is the steady-state arc conductance, i f is the disturbance current, u st is the steady-state arc voltage, u0 is the characteristic arc voltage, and r0 is the characteristic arc resistance.

[0093] In order to build a complete arc fault detection and analysis model, it is necessary to jointly analyze the distribution network model and the disturbance model (including the non-arc model and the arc model). Applying the non-arc disturbance general RL model and the arc fault dynamic resistance model to the distribution network model can respectively build the non-arc disturbance analysis model and the arc disturbance analysis model. The equivalent circuit model is as follows: Figure 5 As shown, Figure 5 (a) is the non-arc disturbance analysis model. Figure 5 (b) is the arc disturbance analysis model.

[0094] The non-arcing disturbance analysis model in S2 is:

[0095]

[0096] Where, R is the static non-arcing resistance, L is the static non-arcing inductance;

[0097] The arc disturbance analysis model is:

[0098]

[0099] Among them, R0 is the constant resistance of the arc.

[0100] When disturbances are detected from the voltage and current information at the slave station, it is necessary to identify arcs from a large number of normal or abnormal disturbances. For a certain disturbance, there are two possibilities: one is an arc disturbance; the other is not an arc disturbance. Therefore, this problem can be understood as a binary hypothesis testing problem. For known measured voltage and current, determine which model the measured signal is more consistent with. When it is more consistent, Figure 5 (a) in the figure is non-arcing disturbance, and when it is more consistent with Figure 5 (b) is an arc fault.

[0101] In S3, the fault detection model is used to judge the measurement signal, including the following steps:

[0102] S3-1: Estimate the load impedance on the transmission line based on the station-end voltage and station-end current before the disturbance;

[0103] S3-2: Estimate the load current and estimate the disturbance current in combination with KCL;

[0104] S3-3: Based on the estimated load impedance, load current and disturbance current, the non-arcing disturbance parameters and arc fault parameters are estimated using the lsqcurvefit function respectively;

[0105] S3-4: recalculating more accurate model parameters and disturbance data based on the estimated non-arcing disturbance parameters and arc fault parameters;

[0106] S3-5: Based on more accurate model parameters and disturbance data, the curve fitting error is used to perform binary hypothesis testing on the results to complete the judgment of the measurement signal.

[0107] First, we need to calculate the load impedance Z load Calculate. Because the equivalent impedance of the line is Z eq Compared to Z load The value of is very small, so here we deal with Z load The influence of line impedance can be ignored at this time. The load impedance parameter R load and L load The station voltage and current before the disturbance can be directly used for estimation.

[0108] S3-1 medium load impedance Z load Estimated to be:

[0109]

[0110] Among them, R load and L load is the load impedance parameter, j is the imaginary part of the load impedance, and f is the grid frequency.

[0111] In the right Figure 5 Before identifying the other model parameters in the above example, the disturbance current i should be accurately estimated. f Since the actual sampling data are all discrete data points, v is used in the calculation s (n) and i s (n) represents the value of the station voltage and current at time nΔt, where n = 1, 2, 3, ...,. For the disturbance current, the station current i s (n) and load current i load (n) Combined with KCL, since the influence of line impedance is ignored, the voltage and current data during the fault period can be used to write a differential equation containing only load parameters, from which the load current can be estimated.

[0112] Estimation of load current in S3-2 load (n) is:

[0113]

[0114] Among them, v s (n) is the value of the station voltage at time nΔt, Δt is the sampling time interval, n is a positive integer, i load (n-1) is the load current at the previous moment;

[0115]

[0116] Among them, i load is the load current;

[0117] The disturbance current is estimated as:

[0118] i f (n) = i s (n)-i load (n)

[0119] Among them, i f (n) is the value of the disturbance current at time nΔt, i s (n) is the value of the station current at time nΔt;

[0120] The differential equation of the disturbance current is:

[0121]

[0122] The station voltage v is known s (n), station-side current i s (n) and the disturbance current i f (n), for Figure 5 (a) in the above also requires an estimate of R eq , L eq , R and L, a total of four unknown parameters, for Figure 5 (b) also requires an estimate of R eq , L eq , R0 and R arc , where R arc Including τ, u0 and r0, a total of six unknown parameters, so this problem can be considered as a composite hypothesis testing problem. When the disturbance is non-arc disturbance, the corresponding Figure 5 (a) in the above equation satisfies the non-arc disturbance analysis model formula, which is assumed to be H0. When the disturbance is an arc fault, the corresponding Figure 5 (b) in the figure satisfies the arc disturbance analysis model formula and is set as H1 assumption.

[0123] For non-arcing disturbances, assuming H0, the unknown parameter vector is expressed as: x0 = (R eq ,L eq ,R,L), known parameters data(n)=(i s (n),di s / dt,i f (n),di f (n) / dt), v s (n)), can be represented by unknown parameter vectors and known parameters, and the function is defined as F0. Taking into account that when performing iterative calculations, errors will be introduced due to the setting of initial conditions, including the disturbance current and the differential calculation of the disturbance current. However, as the number of calculation points increases, the error will decrease. To ensure the accuracy of the detection results, the first n0-1 points with large errors are discarded. When the number of sampling points is N, n0=N / 8 is selected. Finally, based on the idea of ​​least squares, it can be expressed as the following optimization solution problem.

[0124] In S3-3, the lsqcurvefit function is used to estimate the non-arcing disturbance parameters and arc fault parameters respectively, including:

[0125] Estimate the unknown parameters under the assumption of non-arcing disturbance based on the least squares method:

[0126]

[0127] in, is the estimated value of the non-arcing disturbance parameter, n0 is the sampling start time, N is the number of sampling points, F0(·) is the estimated station voltage under the non-arcing disturbance assumption, x is the non-arcing disturbance model parameter, and data(n) is the known parameter vector in the non-arcing disturbance;

[0128] Based on the lsqcurvefit function, the non-arcing disturbance parameter estimation results are obtained

[0129] Estimate the unknown parameters under the assumption of arc fault based on the least squares method:

[0130]

[0131] in, is the estimated value of arc fault parameters, F1(·) is the estimated station terminal voltage under the assumption of arc fault;

[0132] The arc fault model is:

[0133]

[0134] Wherein, g(n) is the arc conductance at time nΔt, and g(n-1) is the arc conductance at the previous time;

[0135] The arc conductance g(n) is:

[0136]

[0137] Based on the lsqcurvefit function, the arc fault parameter estimation results are obtained

[0138] The superscript ∧ indicates an estimated value.

[0139] If g(1)=1, g(n) can be used (i f (n),τ,u0,r0). The unknown parameter vector is expressed as: x1=(R eq ,L eq ,R0,τ,u0,r0), known parameters data(n)=(i s (n),di s / dt,i f (n)), v s (n) can also be expressed by unknown parameter vectors and known parameters, and the function is defined as F1. At the same time, the errors that may be introduced by iteration, initial value, disturbance current and arc conductance are also taken into consideration. The first n0-1 points are also discarded, and the unknown parameters under the H1 assumption are still estimated based on the least squares method.

[0140] Because in the formula When estimating the load impedance, the voltage phasor and current phasor at the station end are used. The impedance obtained at this time should be the sum of the load impedance and the line equivalent impedance. The R eq With L eq After that, a more accurate load impedance Z Load It can be calculated by the following formula:

[0141] Z Load =Z load -Z eq

[0142] Then the load current, disturbance current, disturbance current differential and arc conductance are recalculated to obtain more accurate disturbance data.

[0143] After obtaining the model parameters and the data during the fault period, the results are tested for binary hypothesis. At this time, the parameters are reversed into the hypothesized model, and the disturbance type is determined based on the difference between the measured data and the hypothesized model, using the curve fitting error as the test standard.

[0144] The curve fitting error in S3-5 includes the curve fitting error under the non-arcing disturbance assumption and the curve fitting error under the arc fault assumption;

[0145] The curve fitting error e under the assumption of no arc disturbance non-arc for:

[0146]

[0147] According to the non-arcing disturbance parameter estimation results Calculate the station voltage under the assumption of no arc disturbance for:

[0148]

[0149] The curve fitting error e under the arc fault assumption condition arc for:

[0150]

[0151] According to the arc fault parameter estimation results Calculation of the station voltage under the assumption of an arc fault for:

[0152]

[0153] In S3-5, the curve fitting error is used to perform a binary hypothesis test on the results. The formula is:

[0154]

[0155] Among them, H0 represents non-arcing disturbance and H1 represents arc fault.

[0156] When the fitting error under the non-arc disturbance assumption is small, it is judged as a non-arc disturbance; when the fitting error under the arc fault assumption is small, it is judged as an arc disturbance. When both fitting errors are too large, the fitting accuracy is considered to be low, so it is not used as a basis for fault detection.

[0157] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the invention.

Claims

1. A distribution network arc fault detection method based on power disturbance modeling, characterized in that: The following steps are involved: S1: Build a distribution network model, and build a disturbance model based on the distribution network model; S2: Apply the disturbance model to the distribution network model, and build a non-arc disturbance analysis model and an arc disturbance analysis model respectively; S3: Based on the non-arcing disturbance analysis model and the arc disturbance analysis model, the fault detection model is used to judge the measurement signal to complete the distribution network arc fault detection based on power disturbance modeling.

2. The method for detecting arc faults in distribution networks based on power disturbance modeling according to claim 1, characterized in that: The network distribution model in S1 is: Among them, v s is the station voltage, i s is the station current, R eq is the equivalent resistance between the substation and the disturbance current, L eq is the equivalent inductance between the substation and the disturbance, v f is the voltage at the disturbance point, and t is the time.

3. The method for detecting arc faults in distribution networks based on power disturbance modeling according to claim 2, characterized in that: The disturbance model in S1 includes a non-arc disturbance model and an arc fault model; The non-arc disturbance model adopts a non-arc disturbance general RL model; The arc fault model is: Where g is the dynamic arc conductance, τ is the time constant of the arc, G is the steady-state arc conductance, i f is the disturbance current, u st is the steady-state arc voltage, u0 is the characteristic arc voltage, and r0 is the characteristic arc resistance.

4. The method for detecting arc faults in distribution networks based on power disturbance modeling according to claim 3 is characterized in that: The non-arc disturbance analysis model in S2 is: Where, R is the static non-arcing resistance, L is the static non-arcing inductance; The arc disturbance analysis model is: Among them, R0 is the constant resistance of the arc.

5. The method for detecting arc faults in distribution networks based on power disturbance modeling according to claim 4, characterized in that: In S3, the fault detection model is used to judge the measurement signal, including the following sub-steps: S3-1: Estimate the load impedance on the transmission line based on the station-end voltage and station-end current before the disturbance; S3-2: Estimate the load current and estimate the disturbance current in combination with KCL; S3-3: Based on the estimated load impedance, load current and disturbance current, the non-arcing disturbance parameters and arc fault parameters are estimated using the lsqcurvefit function respectively; S3-4: recalculating more accurate model parameters and disturbance data based on the estimated non-arcing disturbance parameters and arc fault parameters; S3-5: Based on more accurate model parameters and disturbance data, the curve fitting error is used to perform binary hypothesis testing on the results to complete the judgment of the measurement signal.

6. The method for detecting arc faults in distribution networks based on power disturbance modeling according to claim 5, characterized in that: The load impedance Z in S3-1 load Estimated to be: Among them, R load and L load is the load impedance parameter, j is the imaginary part of the load impedance, and f is the grid frequency.

7. The method for detecting arc faults in distribution networks based on power disturbance modeling according to claim 6, characterized in that: The load current estimation i in S3-2 load (n) is: Among them, v s (n) is the value of the station voltage at time nΔt, Δt is the sampling time interval, n is a positive integer, i load (n-1) is the load current at the previous moment; Among them, i load is the load current; The disturbance current is estimated as: i f (n)=i s (n)-i load (n) Among them, i f (n) is the value of the disturbance current at time nΔt, i s (n) is the value of the station current at time nΔt; The differential equation of the disturbance current is:

8. The method for detecting arc faults in distribution networks based on power disturbance modeling according to claim 7, characterized in that: In S3-3, the non-arcing disturbance parameters and arc fault parameters are estimated respectively using the lsqcurvefit function, including: Estimate the unknown parameters under the assumption of non-arcing disturbance based on the least squares method: in, is the estimated value of the non-arcing disturbance parameter, n0 is the sampling start time, N is the number of sampling points, F0(·) is the estimated station voltage under the non-arcing disturbance assumption, x is the non-arcing disturbance model parameter, and data(n) is the known parameter vector in the non-arcing disturbance; Based on the lsqcurvefit function, the non-arcing disturbance parameter estimation results are obtained Estimate the unknown parameters under the assumption of arc fault based on the least squares method: in, is the estimated value of arc fault parameters, F1(·) is the estimated station terminal voltage under the assumption of arc fault; The arc fault model is: Wherein, g(n) is the arc conductance at time nΔt, and g(n-1) is the arc conductance at the previous time; The arc conductance g(n) is: Based on the lsqcurvefit function, the arc fault parameter estimation results are obtained The superscript ∧ indicates an estimated value.

9. The method for detecting arc faults in distribution networks based on power disturbance modeling according to claim 8, characterized in that: The curve fitting error in S3-5 includes the curve fitting error under the non-arc disturbance assumption condition and the curve fitting error under the arc fault assumption condition; The curve fitting error e under the assumption of no arc disturbance non-arc for: According to the non-arcing disturbance parameter estimation results Calculate the station voltage under the assumption of no arc disturbance for: The curve fitting error e under the arc fault assumption condition arc for: According to the arc fault parameter estimation results Calculation of the station voltage under the assumption of an arc fault for:

10. The method for detecting arc faults in distribution networks based on power disturbance modeling according to claim 9, characterized in that: In S3-5, the curve fitting error is used to perform a binary hypothesis test on the results, and the formula is: Among them, H0 represents non-arcing disturbance and H1 represents arc fault.