Power distribution network fault positioning method, device, equipment and medium

Abstract

The invention discloses a power distribution network fault positioning method and device, equipment and a medium. The method comprises the following steps: S1, collecting multi-source data; s2, correcting a line impedance value by the meteorological attenuation factor; s3, outputting a health score of the switch equipment; s4, calculating a node fault confidence coefficient; s5, searching a fault interval; s6, generating an inspection instruction; s7, performing iterative training; according to the invention, the meteorological attenuation factor is introduced to correct the line impedance measurement value in real time, and the health score is combined to effectively solve the fault missed judgment problem caused by rainstorm or strong wind; an inspection instruction is generated according to the node fault confidence of the fault interval, the overall fault positioning time is shortened, and the power supply recovery speed is increased; the machine account and real-time data are fused, equipment state early warning is realized, high-risk equipment is accurately identified, expansion of a fault interval caused by false triggering of an old switch and participation of historical maintenance times in equipment health scoring are avoided, and the fault rate of a pilot line is effectively reduced.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of power systems, in particular to a power distribution network fault positioning method, device, equipment and medium. BACKGROUND

[0002] The power distribution network refers to the part of the power system directly supplying power to users, including all devices and lines between the transformer substation and the users; the main function of the power distribution network is to efficiently and reliably distribute power to each user; the voltage level of the power distribution network is usually low, and the power distribution network fault refers to various problems occurring in the operation process of the power distribution network, resulting in power supply interruption or instability; the current power distribution network fault positioning still has certain deficiencies:

[0003] Firstly, the traditional traveling wave method has high signal attenuation rate in thunderstorm and gale weather, the impedance method is affected by the coupling of load fluctuation and weather, the positioning error is large, and the fault misjudgment rate is high;

[0004] Secondly, the existing SCADA system only monitors real-time electrical quantities, lacks dynamic evaluation of switch aging and maintenance history, part of the misjudgment is caused by abnormal old equipment, and manual inspection is time-consuming;

[0005] Finally, the unmanned aerial vehicle and manual inspection operate independently, lack intelligent cooperation, in extreme weather such as typhoon, the manual inspection range of the fault interval is large, and the power supply recovery appears long time delay;

[0006] Therefore, it is necessary to design a power distribution network fault positioning method, device, equipment and medium. SUMMARY

[0007] The application aims to provide a power distribution network fault positioning method, device, equipment and medium, to solve the problems in the prior art that the traditional traveling wave method has high signal attenuation rate in thunderstorm and gale weather, the impedance method is affected by the coupling of load fluctuation and weather, the positioning error is large, and the fault misjudgment rate is high, and the existing SCADA system only monitors real-time electrical quantities, lacks dynamic evaluation of switch aging and maintenance history, part of the misjudgment is caused by abnormal old equipment, and manual inspection is time-consuming, and the unmanned aerial vehicle and manual inspection operate independently, lack intelligent cooperation, in extreme weather such as typhoon, the manual inspection range of the fault interval is large, and the power supply recovery appears long time delay.

[0008] To achieve the above-mentioned purpose, the application provides the following technical scheme:

[0009] In a first aspect, a power distribution network fault positioning method is provided, comprising the following steps:

[0010] S1, collecting real-time electrical quantity data of each node of the power distribution network through the SCADA system, collecting meteorological data through the meteorological monitoring station, and obtaining the account data and real-time monitoring data of the switch device from the device management database, the account data including the brand code of the switch device, the service life A and the historical maintenance number N repair , and the real-time monitoring data including the switch temperature T and the operation number N ops ;

[0011] S2, when the rainfall ε or the wind speed v in the meteorological data exceeds the threshold value, calculating the meteorological attenuation factor α(ε, v) and correcting the line impedance value Z' measured by the SCADA system, and calculating the current mutation ΔI in the line i ;

[0012] S3, inputting the account data and the real-time monitored switch temperature T and operation number N ops into the LSTM time series neural network model, and outputting the switch device health score H score ;

[0013] S4, determining the switch device weight w score according to the device health score H i , combining the current mutation ΔI in the electrical quantity data i , and calculating the node fault confidence C fault (i);

[0014] S5, constructing a power distribution network topology graph taking the switch device as a node and the line as an edge, taking the node fault confidence C fault (i) as the node weight, and searching for the path with the largest cumulative weight as the fault interval through the reverse Dijkstra algorithm;

[0015] S6, generating an inspection instruction according to the node fault confidence C fault (i) of each node in the fault interval;

[0016] S7, updating the historical database based on the actual fault positioning result, and iteratively training the LSTM time series neural network model and the meteorological attenuation factor α(ε, v) monthly.

[0017] As a further technical solution of the application, in step S2, the calculation formula of the meteorological attenuation factor α(ε, v) is:

[0018]

[0019] Wherein, ε is the rainfall, v is the wind speed, k1 is the rainfall compensation coefficient, and k2 is the wind speed compensation coefficient.

[0020] As a further technical solution of the present application, the rainfall compensation coefficient k1 and the wind speed compensation coefficient k2 are obtained by fitting historical thunderstorm weather data, and the optimization objective function is:

[0021]

[0022] ΔZ j is a historical impedance measurement value, α(ε j ,v j ) is a historical meteorological attenuation factor, and Z raw,j is a historical SCADA original measurement impedance.

[0023] As a further technical solution of the present application, in the step S2, the correction formula of the line impedance value Z' is:

[0024] Z′=Z raw ×α(ε,v)

[0025] Wherein, Z' is the corrected line impedance value, Z raw is the original SCADA system measurement line impedance, and α(ε,v) is the meteorological attenuation factor.

[0026] As a further technical solution of the present application, in the step S2, the calculation of the current mutation ΔI i in the line is:

[0027] According to the corrected line impedance value Z' and the node voltage U i , the expected current under normal working condition is calculated:

[0028]

[0029] Take the absolute difference between the SCADA measured current I i and the expected current:

[0030] ΔI i =|I i -I normal,i |

[0031] Wherein, I normal,i is the expected current under normal working condition.

[0032] As a further technical solution of the present application, the calculation formula of the LSTM time sequence neural network model output health score H score in the step S3 is:

[0033] Input feature preprocessing, static feature vector:

[0034]

[0035] Wherein, A is the service life of the switch device, Amax is the maximum service life of the switch device, N repair is the historical maintenance times of the switch device, N repair,max is the maximum maintenance times allowed by the switch device, b is the encoding vector of the switch device, is the matrix transpose symbol;

[0036] Dynamic feature matrix:

[0037]

[0038] wherein, t w is the length of the time window, T t is the temperature of the switch device at the tth sampling point, T rated is the rated operating temperature of the switch device, N ops,t is the cumulative operation times up to the tth sampling point, N max is the maximum operation times allowed by the switch device;

[0039] The LSTM time series neural network model processes time series, and for each time step t (t = 1, 2,..., t w , the following are calculated in order:

[0040] Forget gate:

[0041] f t = σ (W f · [h t-1 ; d t ] + b f )

[0042] Input gate:

[0043] i t = σ (W i · [h t-1 ; d t ] + b i )

[0044] Candidate memory:

[0045]

[0046] Updated memory:

[0047]

[0048] Output gate:

[0049] o t = σ (W o · [h t-1 ; d t ] + b o )

[0050] Hidden state:

[0051] h t = o t ⊙tanh(C t )

[0052] where h t-1 is the hidden state of the previous time step, C t-1 is the memory state of the previous time step, [h t-1 ; d t ] concatenates the vectors h t-1 and d t , sigma is the Sigmoid activation function, tanh is the hyperbolic tangent activation function, and is the element-wise multiplication, W f , W i , W C , W o are weight matrices, and b f , b i , b C , b o are bias vectors, and the final hidden state of the last time step is obtained

[0053] Feature fusion and score output:

[0054]

[0055] where W1, b1 are the weight matrix and bias vector of the first fully connected layer, W2, b2 are the weight matrix and bias vector of the second fully connected layer, and is the rectified linear unit activation function.

[0056] As a further technical solution of the present application, the weight distribution rule in step S4 is:

[0057] Health score H score ≥ 0.8, switch device weight w i = 1.0;

[0058] 0.5≤ health score H score < 0.8, switch device weight w i = 0.6;

[0059] Health score H score < 0.5, switch device weight w i = 0.2.

[0060] As a further technical solution of the present application, in step S4, the calculation formula of node fault confidence C fault (i) is:

[0061]

[0062] wherein, w i is a switch device health weight, ΔI i is a current mutation value, I max is a maximum allowable current of the line.

[0063] As a further technical solution of the present application, in the step S5, the reverse Dijkstra algorithm search path weight calculation formula is:

[0064] W path =∑ i∈path C fault (i)

[0065] The path with the largest cumulative weight is taken as the fault interval.

[0066] As a further technical solution of the present application, the inspection instruction in the step S6 includes dispatching a drone or manual line inspection, wherein the drone is equipped with an infrared imager and a visible light camera, and transmits back images in real time, and the node fault confidence C fault (i)≥0.7 is a high confidence area, and the drone is preferentially scanned quickly, and 0.4≤node fault confidence C fault (i)<0.7 is a medium confidence area, and the nearest line inspection team is accurately dispatched.

[0067] In a second aspect, a power system device is provided, comprising a data acquisition module, a data processing module and an instruction module, the data acquisition module being capable of acquiring multi-source data; the data processing module being capable of performing data processing of the acquired data according to steps S2 to S5, and the instruction module being capable of giving a treatment instruction according to the data processing result in step S6.

[0068] In a third aspect, an electronic device is provided, comprising a processor, a memory and a computer program stored on the memory, the computer program being executed by the processor to implement the power distribution network fault locating method in the first aspect.

[0069] In a fourth aspect, a computer readable medium is provided, having a computer program stored thereon, the computer program being executed by a processor to implement the power distribution network fault locating method in the first aspect.

[0070] Compared with the prior art, the power distribution network fault locating method, device, equipment and medium have the following beneficial effects:

[0071] By introducing the meteorological attenuation factor α(ε,v) to real-time correct the line impedance measurement value Z', combining the health score weighting mechanism, the positioning error is compressed, the positioning success rate is improved, and the fault missed judgment problem caused by heavy rain or strong wind is effectively solved;

[0072] According to the node fault confidence Cfault (i), generating a patrol instruction according to the node fault confidence C fault (i) selecting unmanned aerial vehicles or manual line inspection, shortening response time, and reducing labor input, overall fault positioning time is shortened, and power supply recovery speed is improved;

[0073] The LSTM time sequence neural network model fuses the account book and real-time data, realizes early warning of device state, accurately identifies high-risk devices, and automatically reduces weight, avoids mis-triggering of old switches to expand fault intervals, and historical maintenance times N repair Participate in equipment health score H score , guide preventive replacement decision, and effectively reduce the fault rate of the pilot line. BRIEF DESCRIPTION OF DRAWINGS

[0074] Figure 1 The method flowchart of the present application is shown. DETAILED DESCRIPTION

[0075] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0076] Please refer to the accompanying Figure 1 The present application provides a power distribution network fault positioning method, which comprises the following steps:

[0077] S1, collecting real-time electrical quantity data of each node of the power distribution network through the SCADA system, collecting meteorological data through the meteorological monitoring station, and obtaining account book data and real-time monitoring data of the switch device from the equipment management database, the account book data including switch device brand code, service life A and historical maintenance times N repair , the real-time monitoring data including switch temperature T, operation times N ops ;

[0078] S2, when the rainfall ε or wind speed v in the meteorological data exceeds the threshold value, calculate the meteorological attenuation factor α(ε, v), correct the line impedance value Z' measured by the SCADA system, and calculate the current mutation ΔI i ;

[0079] The calculation formula of the meteorological attenuation factor α(ε, v) is:

[0080]

[0081] Wherein, ε is the rainfall, v is the wind speed, k1 is the rainfall compensation coefficient, and k2 is the wind speed compensation coefficient.

[0082] The rainfall compensation coefficient k1 and the wind speed compensation coefficient k2 are obtained by fitting historical thunderstorm weather data, and the optimization objective function is:

[0083]

[0084] Where, ΔZ j is a historical impedance measurement value, α(ε j ,v j ) is a historical meteorological attenuation factor, Z raw,j is a historical SCADA original measurement impedance;

[0085] The correction formula of the line impedance value Z' is:

[0086] Z' = Z raw × α(ε,v)

[0087] Where, Z' is the corrected line impedance value, Z raw is the SCADA system original measurement line impedance, and α(ε,v) is the meteorological attenuation factor;

[0088] The calculation of the current mutation ΔI i in the line is:

[0089] According to the corrected line impedance value Z' and the node voltage U i , the expected current under normal working conditions is calculated:

[0090]

[0091] Take the absolute difference between the SCADA measured current I i and the expected current:

[0092] ΔI i = |I i -I normal,i |

[0093] Where, I normal,i is the expected current under normal working conditions;

[0094] S3, input the ledger data and the real-time monitored switch temperature T, operation times N ops into the LSTM time series neural network model, and output the switch device health score H score ;

[0095] The calculation formula of the health score H score output by the LSTM time series neural network model is:

[0096] Input feature preprocessing, static feature vector:

[0097]

[0098] Where A represents the service life of the switchgear, A max For the maximum service life of the switchgear, N repair N represents the number of historical maintenance visits to the switchgear. repair,max Let b be the maximum number of maintenance cycles allowed for the switchgear, and b be the switchgear encoding vector. This is the matrix transpose symbol;

[0099] Dynamic feature matrix:

[0100]

[0101] Among them, t w T is the length of the time window. t Let T be the temperature of the switching equipment at the t-th sampling point. rated The rated operating temperature of the switchgear, N ops,t N represents the cumulative number of operations up to the t-th sampling point. max The maximum number of operations allowed for the switching equipment;

[0102] Temporal processing of LSTM temporal neural network models, for each time step t (t=1,2,...,t) w ), calculate in order:

[0103] Forgotten Gate:

[0104] f t =σ(W f ·[h t-1 ;d t ]+b f )

[0105] Input Gate:

[0106] i t =σ(W i ·[h t-1 ;d t ]+b i )

[0107] Candidate memories:

[0108]

[0109] Update memory:

[0110]

[0111] Output gate:

[0112] o t =σ(W o ·[h t-1 ;dt ]+b o )

[0113] Hidden state:

[0114] h t =o t ⊙tanh(C t )

[0115] where h t-1 is the hidden state of the previous time step, C t-1 is the memory state of the previous time step, [h t-1 ; d t ] concatenates the vectors h t-1 and d t , σ is the Sigmoid activation function, tanh is the hyperbolic tangent activation function, ⊙ is the element-wise multiplication, W f , W i , W C , W o are weight matrices, b f , b i , b C , b o are bias vectors, and the final hidden state of the last time step is obtained

[0116] Feature fusion and score output:

[0117]

[0118] where W1, b1 are the weight matrix and bias vector of the first fully connected layer, W2, b2 are the weight matrix and bias vector of the second fully connected layer, and ReLU is the rectified linear unit activation function.

[0119] S4, determine the switch device weight w score according to the device health score H i , combined with the current mutation ΔI i in the electrical quantity data, calculate the node fault confidence C fault (i);

[0120] The weight allocation rule is:

[0121] Health score H score ≥ 0.8, switch device weight w i = 1.0;

[0122] 0.5≤ health score H score <0.8, switch device weight w i = 0.6;

[0123] Health score H score<0.5, switch device weight w i = 0.2;

[0124] Node fault confidence C fault The calculation formula of (i) is:

[0125]

[0126] Wherein, w i is the switch device health weight, ΔI i is the current mutation value, I max is the maximum allowable current of the line;

[0127] S5, build the power distribution network topology graph with switch device as node and line as edge, with node fault confidence C fault (i) as node weight, search the path with the largest cumulative weight as the fault interval by reverse Dijkstra algorithm;

[0128] The path weight calculation formula of reverse Dijkstra algorithm is:

[0129] W path = ∑ i∈patn C fault (i)

[0130] The path with the largest cumulative weight is taken as the fault interval;

[0131] S6, according to the node fault confidence C fault (i) of each node in the fault interval, generate inspection instruction;

[0132] The inspection instruction includes scheduling unmanned aerial vehicle or manual line inspection, wherein the unmanned aerial vehicle is equipped with infrared imager and visible light camera, and transmits back image in real time, the node fault confidence C fault (i)≥0.7 is a high confidence area, and the unmanned aerial vehicle is preferentially scanned quickly, 0.4≤the node fault confidence C fault (i)<0.7 is a medium confidence area, and the nearest line inspection team is accurately scheduled;

[0133] S7, update the historical database based on the actual fault positioning result, and iteratively train the LSTM time sequence neural network model and the meteorological attenuation factor alpha (epsilon, v) per month.

[0134] An embodiment provided by the application: a power system device, comprising a data acquisition module, a data processing module and an instruction module, the data acquisition module can acquire multi-source data;The data processing module can perform data processing of steps S2 to S5 on the collected data, and the instruction module can give treatment instruction according to the data processing result in step S6.

[0135] An embodiment provided by the present application: an electronic device, comprising a processor, a memory and a computer program stored on the memory, the computer program being executed by the processor to implement a power distribution network fault location method.

[0136] An embodiment provided by the present application: a computer readable medium having a computer program stored thereon, the computer program being executed by a processor to implement a power distribution network fault location method.

[0137] In summary, the power distribution network fault location method, device, equipment and medium provided by the present application are characterized in that:

[0138] The weather attenuation factor alpha (ε, v) is introduced to correct the line impedance measurement value Z' in real time, the positioning error is compressed by combining the health score weighting mechanism, the positioning success rate is improved, and the problem of fault misjudgment caused by heavy rain or strong wind is effectively solved;

[0139] According to the node fault confidence C fault (i) of each node in the fault interval, the inspection instruction is generated according to the node fault confidence C fault (i), the unmanned aerial vehicle or manual line inspection is selected to shorten the response time and reduce the labor input, the overall fault location time is shortened, and the power supply recovery speed is improved;

[0140] The LSTM time sequence neural network model fusing the account book and real-time data realizes the early warning of the equipment state, accurately identifies the high-risk equipment, and automatically reduces the weight, avoids the expansion of the fault interval caused by the false triggering of the old switch, and the historical maintenance times N repair participate in the equipment health score H score , guide the preventive replacement decision, and effectively reduce the fault rate of the pilot line.

[0141] It is obvious to those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application. Therefore, the embodiments should be regarded as exemplary and non-limiting, the scope of the present application is defined by the appended claims rather than the above description, and therefore all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present application. Any reference signs in the claims should not be regarded as limiting the claims involved.

Claims

1. A method for locating faults in a power distribution network, characterized in that: Includes the following steps: S1. Collect real-time electrical quantity data of each node in the distribution network through the SCADA system, collect meteorological data through the meteorological monitoring station, and obtain the ledger data and real-time monitoring data of the switchgear from the equipment management database. The ledger data includes the brand code of the switchgear, the service life A, and the number of historical maintenance N. repair Real-time monitoring data includes switch temperature T and number of operations N. ops ; S2. When the rainfall ε or wind speed v in the meteorological data exceeds the threshold, calculate the meteorological attenuation factor α(ε,v), correct the line impedance value Z′ measured by the SCADA system, and calculate the current surge ΔI in the line. i ; S3. Combine the ledger data with the real-time monitored switch temperature T and number of operations N. ops Input an LSTM temporal neural network model and output a switching device health score H. score ; S4. Based on the equipment health score H score Determine the weight w of the switching equipment i Combined with the current surge ΔI in the electrical quantity data i Calculate the node fault confidence C fault (i); S5. Construct a distribution network topology with switching equipment as nodes and lines as edges, and use node fault confidence C. fault (i) represents the node weight, and the path with the largest cumulative weight is searched using the reverse Dijkstra algorithm as the fault interval; S6. Based on the node fault confidence C of each node within the fault interval. fault (i) Generate inspection instructions; S7. Update the historical database based on the actual fault location results, and iteratively train the LSTM time-series neural network model and the meteorological attenuation factor α(ε,v) on a monthly basis.

2. The method for locating faults in a power distribution network according to claim 1, characterized in that: In step S2, the formula for calculating the meteorological attenuation factor α(ε,v) is: Where ε is the rainfall, v is the wind speed, k1 is the rainfall compensation coefficient, and k2 is the wind speed compensation coefficient.

3. The method for locating faults in a power distribution network according to claim 2, characterized in that: The rainfall compensation coefficient k1 and wind speed compensation coefficient k2 were obtained by fitting historical thunderstorm weather data, and the optimization objective function is: Where, ΔZ j The historical impedance measurement value, α(ε) j ,v j Z is the historical meteorological attenuation factor. raw,j This is the original measurement impedance of the historical SCADA system.

4. The method for locating faults in a power distribution network according to claim 1, characterized in that: In step S2, the correction formula for the line impedance value Z′ is: Z′=Z raw ×α(ε,v) Where Z′ is the corrected line impedance value, Z raw The original measured line impedance of the SCADA system is α(ε,v), which is the meteorological attenuation factor.

5. The method for locating faults in a power distribution network according to claim 1, characterized in that: In step S2, the current in the circuit changes abruptly by ΔI. i The calculation is as follows: Based on the corrected line impedance value Z′ and node voltage U i Calculate the expected current under normal operating conditions: Take the actual measured current I from SCADA i The absolute difference from the expected current: ΔI i =|I i -I normal,i | Among them, I normal,i This is the expected current under normal operating conditions.

6. The method for locating faults in a power distribution network according to claim 1, characterized in that: The LSTM temporal neural network model in step S3 outputs a health score H. score The formula for calculation is: Input feature preprocessing, static feature vector: Where A represents the service life of the switchgear, A max For the maximum service life of the switchgear, N repair N represents the number of historical maintenance visits to the switchgear. repair,max Let b be the maximum number of maintenance cycles allowed for the switchgear, and b be the switchgear encoding vector. This is the matrix transpose symbol; Dynamic feature matrix: Among them, t w T is the length of the time window. t Let T be the temperature of the switching equipment at the t-th sampling point. rated The rated operating temperature of the switchgear, N ops,t N represents the cumulative number of operations up to the t-th sampling point. max The maximum number of operations allowed for the switching equipment; Temporal processing of LSTM temporal neural network models, for each time step t (t=1,2,...,t) w ), calculate in order: Forgotten Gate: f t =σ(W f ·[h t-1 ;d t ]+b f ) Input Gate: i t =σ(W i ·[h t-1 ;d t ]+b i ) Candidate memories: Update memory: Output gate: the t =σ(W o ·[h t-1 ;d t ]+b o ) Hidden state: h t =o t ⊙tanh(C t ) Among them, h t-1 C is the hidden state of the previous time step. t-1 The memory state of the previous time step, [h t-1 ;d t ]Transfer vector h t-1 and d t Concatenation, σ is the Sigmoid activation function, tanh is the hyperbolic tangent activation function, ⊙ is element-wise multiplication, W f W i W C W o Let b be the weight matrix. f b i b C b o Using the bias vector, we can ultimately obtain the hidden state at the last time step. Feature fusion and scoring output: Where W1 and b1 are the weight matrix and bias vector of the first fully connected layer, W2 and b2 are the weight matrix and bias vector of the second fully connected layer, and b1 is the modified linear unit activation function.

7. The method for locating faults in a distribution network according to claim 1, characterized in that: The weight allocation rule in step S4 is as follows: Health Score H score ≥0.8, switchgear weight w i =1.0; 0.5≤Health Score H score <0.8, switchgear weight w i =0.6; Health Score H score <0.5, switchgear weight w i =0.

2.

8. The method for locating faults in a power distribution network according to claim 1, characterized in that: In step S4, the node fault confidence C fault The formula for (i) is: Among them, w i For the health weight of switching equipment, ΔI i I represents the sudden change in current. max This represents the maximum allowable current for the line.

9. The method for locating faults in a power distribution network according to claim 1, characterized in that: In step S5, the formula for calculating the path weight of the reverse Dijkstra algorithm search is: The path with the highest cumulative weight is selected as the fault interval.

10. The method for locating faults in a power distribution network according to claim 1, characterized in that: The inspection instructions in step S6 include dispatching drones or manual inspections, wherein the drones are equipped with infrared imagers and visible light cameras to transmit images in real time, and the node fault confidence level C. fault (i) ≥ 0.7 indicates a high-confidence region, which should be prioritized for rapid scanning by UAVs; 0.4 ≤ node fault confidence level C fault (i) <0.7 indicates a medium confidence area, allowing for precise dispatch of the nearest patrol team.

11. A power system device, characterized in that, It includes a data acquisition module, a data processing module, and an instruction module. The data acquisition module can acquire data from multiple sources; the data processing module can process the acquired data in steps S2 to S5; and the instruction module can issue a processing instruction based on the data processing result in step S6.

12. An electronic device comprising a processor, a memory, and a computer program stored in the memory, characterized in that, When the computer program is executed by the processor, it implements the power distribution network fault location method as described in any one of claims 1-10.

13. A computer-readable medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the power distribution network fault location method as described in any one of claims 1-10.

Images