CN113533906B - A kind of intelligent overhead transmission line fault type diagnosis method and system - Google Patents

Abstract

The invention relates to the technical field of fault diagnosis of transmission lines, in particular to a method and system for diagnosing fault types of intelligent overhead transmission lines. The present invention can firstly judge whether it is a lightning strike fault by using the traveling wave characteristics for the installed distributed fault monitoring and diagnosis device system, and if it is a lightning strike fault, use the lightning locating system to correct the lightning fault probability and give the lightning strike fault result. If it is a non-lightning fault, further calculate the fault power frequency characteristics, and then correct the probability of matching the fault cause through the monitoring information of the transmission line corridor such as mountain fire, icing, and weather, and output the diagnosis result. Finally, for systems without distributed fault monitoring and diagnosis devices, directly use the information provided by the dispatching system, such as fault time, reclosing status, fault phase, etc., to calculate the probability corresponding to various fault cause types, and combine the transmission line corridor monitoring information. Make corrections and give diagnostic results. The invention can accurately locate the fault of the transmission line.

Description

A kind of intelligent overhead transmission line fault type diagnosis method and system

technical field

The invention relates to the technical field of fault diagnosis of transmission lines, in particular to a method and system for diagnosing fault types of intelligent overhead transmission lines.

Background technique

At present, the fault diagnosis of transmission lines is mainly based on manual experience, and the elimination method is used to analyze the types of transmission line faults and calculate the location of the fault points one by one. There is a lack of systematic comprehensive analysis and diagnosis technology, especially under the condition of multi-source information, it is impossible to give a clear fault. As a result, the efficiency of fault analysis and search of transmission lines is low, which seriously affects the safe and stable operation of transmission lines.

Transmission line failures mainly include failures caused by lightning strikes, wind deflections, bird flashovers, pollution flashovers, tree flashovers and wildfire failures. The identification of the cause of the failure needs to be based on the understanding of various failure principles and processes, and on this basis, the characteristics of specific failures are mined and analyzed to form the basis for cause identification. Therefore, it is necessary to analyze the failure mechanism of various failure types. analyze. At the same time, the occurrence of faults is related to the operating environment of the transmission line, and the characteristics of different types of faults are different in the recorded wave data. , the characteristics of the non-periodic components of the fault phase current, and the internal factors expressed by the recorded wave data such as transition resistance, to find the characteristic rules, so as to provide a data source for the establishment of the subsequent classification model.

SUMMARY OF THE INVENTION

In order to solve the above problems, the present invention provides a method and system for diagnosing fault types of intelligent overhead transmission lines. The specific technical solutions are as follows:

A fault type diagnosis method for an intelligent overhead transmission line, comprising the following steps:

S1: Judge whether a distributed fault monitoring device or a transmission line corridor environment monitoring device is installed on the faulty system; if the distributed fault monitoring device or transmission line corridor environment monitoring device is installed, go to step S2, otherwise go to step S2 S3;

S2: Search the associated transient traveling wave features, form a combination of feature vectors, input the first support vector machine, and carry out lightning strike and non-lightning strike fault type diagnosis;

S21: If it is a lightning strike fault, query the lightning location system, and combine with the fault location to obtain the lightning current data closest to the fault trip time, and correct the lightning strike fault probability P _lightning . The lightning strike fault probability correction function is f ₁ ;

S22: If it is a non-lightning fault, the system further queries the external online monitoring business platform including the mountain fire monitoring system, the icing system, and the meteorological system, calculates and extracts the features of the power frequency waveform at the time of the failure, obtains the feature combination, and enters the second Support vector machine to get the cause of failure and its probability;

Correct the fault results by combining the data of mountain fire, wind deflection and icing in the multi-source information;

S3: The faulty system is not equipped with distributed fault monitoring devices and transmission line corridor environmental monitoring devices. According to the fault time, reclosing state, and fault phase provided by the dispatching system information, combined with meteorological information, the results can be input into Bayer. The feature combination of the Si network; for the evidence information that has been collected, the prior probability distribution of the corresponding external features of the fault is directly obtained.

Preferably, the correction of the lightning strike failure probability P _lightning in the step S21 is specifically:

P _{lightning-new} =f ₁ (P _{lightning-old} )=a+(1-a)(P _{lightning-old} );

Among them, P _{lightning-new} represents the lightning strike failure probability after correction, P _{lightning-old} represents the lightning strike failure probability before correction, and a is the lightning strike failure probability correction parameter;

The correction parameter a is comprehensively judged according to the amplitude A and the distance D of the nearest lightning current to the fault point found in the lightning location system, and the calculation formula is:

Preferably, it also includes using the corrected lightning strike failure probability in the first vector machine to correct the failure probability of other causes except the lightning strike, as follows:

Among them, P _other-new represents the other-cause failure probability after correction, and P _other-old represents the other-cause failure probability before correction.

Preferably, in the step S22, the mountain fire in the multi-source information is used to correct the fault result, and the details are as follows:

P _fire-new =f ₂ (P _fire-old )=b+(1-b)(P _fire-old );

Among them, P _fire-new represents the corrected wildfire failure probability, P _fire-old represents the wildfire failure probability before correction, f ₂ is the wildfire fault probability correction function, and b is the wildfire fault probability correction parameter;

The calculation method of the correction parameter b of the wildfire failure probability is as follows:

T is the fire temperature.

Preferably, in the step S22, the wind deviation in the multi-source information is used to correct the fault result, and the details are as follows:

P _wind-new =f ₃ (P _wind-old )=c+(1-c)(P _wind-old );

Among them, P _wind-new represents the probability of failure of wind deviation after correction, P _wind-old represents the probability of failure of wind deviation before correction, _f3 is the correction function of probability of failure of wind deviation, and c is the correction parameter of probability of failure of wind deviation;

The calculation method of wind deviation probability correction parameter c is as follows:

c=[min(W,12)+2]/14;

W is the maximum wind speed of the line corridor.

Preferably, in the step S22, icing in the multi-source information is used to correct the fault result, and the details are as follows:

P _ice-new =f ₄ (P _ice-old )=d+(1-d)(P _ice-old );

Among them, P _ice-new represents the icing failure probability after correction, P _ice-old represents the icing failure probability before correction, _f4 is the icing failure probability correction function, and d is the icing failure probability correction parameter;

The calculation method of the icing fault probability correction parameter d is as follows:

d=0.5log(H+1);

H is the maximum ice thickness monitored online.

Preferably, the probability distribution in step S3 includes:

Suppose the fault weather is event A; the reclosing action is event B; the fault phase is event C; the fault month is event D; the fault time is event E; V _i , i=1, 2, 3, 4, 5, represent lightning strike fault, wind deflection fault, bird damage fault, tree flash fault and wildfire fault respectively; the calculation method of fault probability is as follows:

After the failure probability is obtained, the multi-source information is queried again, the same correction function is used to correct the failure probability, and the final result is output.

An intelligent overhead transmission line fault type diagnosis system, comprising a system judgment module, a data module, and a correction module; the system judgment module, the data module, and the correction module are connected in sequence;

The system judging module is used for judging whether a distributed fault monitoring device or a transmission line corridor environment monitoring device is installed for the system in which the fault occurs;

The data module is used for data processing according to the judgment result of the system judgment module, and the details are as follows:

If the system judgment module is equipped with a distributed fault monitoring device and a transmission line corridor environment monitoring device to judge the faulty system, the data module searches for the associated transient traveling wave characteristics, forms a combination of eigenvectors, and inputs the first vector machine to carry out lightning strikes. , Non-lightning fault type diagnosis;

If it is diagnosed as a lightning strike fault, the data module queries the lightning location system, and combines the fault location to obtain the lightning current data closest to the fault trip time, and the correction module corrects the lightning strike fault probability through a correction function;

If it is diagnosed as a non-lightning fault, the data module will further query the external online monitoring business platform including the mountain fire monitoring system, the icing system, and the meteorological system; calculate and extract the features of the power frequency waveform at the time of the fault to obtain the feature combination, and input the second The support vector machine is used to obtain the cause of the failure and its probability; the correction module combines the data of mountain fire, wind deflection and icing in the multi-source information to correct the failure result;

If the system judgment module judges that the faulty system is not equipped with a distributed fault monitoring device and a transmission line corridor environment monitoring device, the data module based on the fault time, reclosing state, and fault phase provided by the dispatching system information, combined with meteorological information, The feature combination that can be input into the Bayesian network is obtained; for the evidence information that has been collected, the prior probability distribution of the corresponding external features of the fault is directly obtained.

Preferably, the correction function is:

P _{lightning-new} =f ₁ (P _{lightning-old} )=a+(1-a)(P _{lightning-old} );

Among them, P _{lightning-new} represents the lightning strike failure probability after correction, P _{lightning-old} represents the lightning strike failure probability before correction, and a is the lightning strike failure probability correction parameter.

The beneficial effects of the present invention are:

The present invention can intelligently identify whether the faulty system is equipped with distributed fault monitoring device, transmission line corridor environment monitoring device, the collected transmission line body data and line corridor environmental monitoring data, and correct and judge the fault type. For the system of the installed distributed fault monitoring and diagnosis device, first use the traveling wave feature to judge whether it is a lightning strike, if it is a lightning strike, use the lightning location system to correct the lightning fault probability and give the lightning strike fault result. If it is not a lightning strike, further calculate the fault power frequency characteristics, and then use the monitoring information of the transmission line corridor such as mountain fire, icing, and meteorology, and then correct the probability of matching the fault cause, and output the diagnosis result. Finally, for systems without distributed fault monitoring and diagnosis devices, directly use the information provided by the dispatching system, such as fault time, reclosing status, and fault phase, to calculate the probabilities corresponding to various fault cause types, and combine the transmission line corridor monitoring information. Make corrections and give diagnostic results. The invention can accurately locate the fault of the transmission line and carry out the analysis of the cause of the tripping fault of the line, which can greatly reduce the workload of line inspection and improve the reliability of power supply.

Description of drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can also be obtained from these drawings without creative effort.

Fig. 1 is the method flow chart of the present invention;

FIG. 2 is a system structure diagram of the present invention.

Detailed ways

In order to better understand the present invention, the present invention will be further described below in conjunction with the accompanying drawings and specific embodiments:

As shown in Figure 1, a method for diagnosing fault types of intelligent overhead transmission lines includes the following steps:

S1: Judge whether a distributed fault monitoring device or a transmission line corridor environment monitoring device is installed on the faulty system; if the distributed fault monitoring device or transmission line corridor environment monitoring device is installed, go to step S2, otherwise go to step S2 S3;

S2: Search the associated transient traveling wave features, form a combination of feature vectors, input the first support vector machine, and carry out lightning strike and non-lightning strike fault type diagnosis;

S21: If it is a lightning strike fault, query the lightning location system, and combine with the fault location to obtain the lightning current data closest to the fault trip time, and correct the lightning strike fault probability P _lightning . The lightning strike fault probability correction function is f ₁ ; P _lightning to amend as follows:

P _{lightning-new} =f ₁ (P _{lightning-old} )=a+(1-a)(P _{lightning-old} );

Among them, P _{lightning-new} represents the lightning strike failure probability after correction, P _{lightning-old} represents the lightning strike failure probability before correction, and a is the lightning strike failure probability correction parameter;

The correction parameter a is comprehensively judged according to the amplitude A and the distance D of the nearest lightning current to the fault point found in the lightning location system, and the calculation formula is:

Since there are many lightning strike faults in the overall data sample, in the first support vector machine, in order to achieve a more ideal classification effect, the penalty coefficient of the lightning strike fault label is controlled. During the parameter adjustment process, the adjusted value is larger, so that more Therefore, it is necessary to correct the probability of the lightning strike according to the amplitude and distance of the lightning current; after correcting the lightning strike probability, in order to make the sum of the probabilities of all causes unique, the other fault causes are corrected proportionally ,details as follows:

Among them, P _other-new represents the other-cause failure probability after correction, and P _other-old represents the other-cause failure probability before correction. S22: If it is a non-lightning fault, the system further queries the external online monitoring business platform including the mountain fire monitoring system, the icing system, and the meteorological system, calculates and extracts the features of the power frequency waveform at the time of the failure, obtains the feature combination, and enters the second Support vector machine to get the cause of failure and its probability;

Correct the fault results by combining the data of mountain fire, wind deflection and icing in the multi-source information;

The fault results are corrected using the wildfires in the multi-source information, as follows:

P _fire-new =f ₂ (P _fire-old )=b+(1-b)(P _fire-old );

Among them, P _fire-new represents the corrected wildfire failure probability, P _fire-old represents the wildfire failure probability before correction, f ₂ is the wildfire fault probability correction function, and b is the wildfire fault probability correction parameter;

The calculation method of the correction parameter b of the wildfire failure probability is as follows:

T is the fire temperature.

The fault results are corrected using the wind bias in the multi-source information, as follows:

P _wind-new =f ₃ (P _wind-old )=c+(1-c)(P _wind-old );

Among them, P _wind-new represents the probability of failure of wind deviation after correction, P _wind-old represents the probability of failure of wind deviation before correction, _f3 is the correction function of probability of failure of wind deviation, and c is the correction parameter of probability of failure of wind deviation;

The calculation method of wind deviation probability correction parameter c is as follows:

c=[min(W,12)+2]/14;

W is the maximum wind speed of the line corridor.

The fault results are corrected by icing in the multi-source information, as follows:

P _ice-new =f ₄ (P _ice-old )=d+(1-d)(P _ice-old );

Among them, P _ice-new represents the icing failure probability after correction, P _ice-old represents the icing failure probability before correction, _f4 is the icing failure probability correction function, and d is the icing failure probability correction parameter;

The calculation method of the icing fault probability correction parameter d is as follows:

d=0.5log(H+1);

H is the maximum ice thickness monitored online.

For uncollected evidence information, if the system fails to find meteorological information at the time of failure, the corresponding probability distribution is 1. S3: The faulty system is not equipped with distributed fault monitoring devices and transmission line corridor environmental monitoring devices. According to the fault time, reclosing state, and fault phase provided by the dispatching system information, combined with meteorological information, the results can be input into Bayer. The feature combination of the Si network; for the evidence information that has been collected, the prior probability distribution of the corresponding external features of the fault is directly obtained. Probability distributions include:

Suppose the fault weather is event A; the reclosing action is event B; the fault phase is event C; the fault month is event D; the fault time is event E; V _i , i=1, 2, 3, 4, 5, represent lightning strike fault, wind deflection fault, bird damage fault, tree flash fault and wildfire fault respectively; the calculation method of fault probability is as follows:

After the failure probability is obtained, the multi-source information is queried again, the same correction function is used to correct the failure probability, and the final result is output.

As shown in Figure 2, an intelligent overhead transmission line fault type diagnosis system includes a system judgment module, a data module, and a correction module; the system judgment module, data module, and correction module are connected in sequence;

The system judging module is used for judging whether a distributed fault monitoring device or a transmission line corridor environment monitoring device is installed for the system in which the fault occurs;

The data module is used for data processing according to the judgment result of the system judgment module, and the details are as follows:

If the system judgment module is equipped with a distributed fault monitoring device and a transmission line corridor environment monitoring device to judge the faulty system, the data module searches for the associated transient traveling wave characteristics, forms a combination of eigenvectors, and inputs the first vector machine to carry out lightning strikes. , Non-lightning fault type diagnosis;

If it is diagnosed as a lightning strike fault, the data module queries the lightning location system, and combines the fault location to obtain the lightning current data closest to the fault trip time, and the correction module corrects the lightning strike fault probability through a correction function;

If it is diagnosed as a non-lightning fault, the data module will further query the external online monitoring business platform including the mountain fire monitoring system, the icing system, and the meteorological system; calculate and extract the features of the power frequency waveform at the time of the fault to obtain the feature combination, and input the second The support vector machine is used to obtain the cause of the failure and its probability; the correction module combines the data of mountain fire, wind deflection and icing in the multi-source information to correct the failure result;

If the system judgment module judges that the faulty system is not equipped with a distributed fault monitoring device and a transmission line corridor environment monitoring device, the data module based on the fault time, reclosing state, and fault phase provided by the dispatching system information, combined with meteorological information, The feature combination that can be input into the Bayesian network is obtained; for the evidence information that has been collected, the prior probability distribution of the corresponding external features of the fault is directly obtained.

The correction function of the correction module for the lightning strike failure probability is:

P _{lightning-new} =f ₁ (P _{lightning-old} )=a+(1-a)(P _{lightning-old} );

Among them, P _{lightning-new} represents the lightning strike failure probability after correction, P _{lightning-old} represents the lightning strike failure probability before correction, and a is the lightning strike failure probability correction parameter. The correction parameter a is comprehensively judged according to the amplitude A and the distance D of the nearest lightning current to the fault point found in the lightning location system. The calculation formula is:

The correction function of the correction module for the probability of wildfire failure is:

P _fire-new =f ₂ (P _fire-old )=b+(1-b)(P _fire-old );

Among them, P _fire-new represents the corrected wildfire failure probability, P _fire-old represents the wildfire failure probability before correction, f ₂ is the wildfire fault probability correction function, and b is the wildfire fault probability correction parameter;

The calculation method of the correction parameter b of the wildfire failure probability is as follows:

T is the fire temperature.

The correction function of the correction module for the probability of wind deviation failure is:

P _wind-new =f ₃ (P _wind-old )=c+(1-c)(P _wind-old );

Among them, P _wind-new represents the probability of failure of wind deviation after correction, P _wind-old represents the probability of failure of wind deviation before correction, _f3 is the correction function of probability of failure of wind deviation, and c is the correction parameter of probability of failure of wind deviation;

The calculation method of wind deviation probability correction parameter c is as follows:

c=[min(W,12)+2]/14;

W is the maximum wind speed of the line corridor.

The correction function of the correction module for the probability of icing failure is:

P _ice-new =f ₄ (P _ice-old )=d+(1-d)(P _ice-old );

Among them, P _ice-new represents the icing failure probability after correction, P _ice-old represents the icing failure probability before correction, _f4 is the icing failure probability correction function, and d is the icing failure probability correction parameter;

The calculation method of the icing fault probability correction parameter d is as follows:

d=0.5log(H+1);

H is the maximum ice thickness monitored online.

The invention provides a fault type diagnosis method and system for an intelligent overhead transmission line. For an installed system of distributed fault monitoring and diagnosis devices, the traveling wave feature is used to determine whether it is a lightning strike, and if it is a lightning strike, a lightning location system is used. Correct its lightning failure probability and give the lightning strike failure result. If it is not a lightning strike, further calculate the fault power frequency characteristics, and then use the monitoring information of the transmission line corridor such as mountain fire, icing, and meteorology, and then correct the probability of matching the fault cause, and output the diagnosis result. Finally, for systems without distributed fault monitoring and diagnosis devices, directly use the information provided by the dispatching system, such as fault time, reclosing status, and fault phase, to calculate the probabilities corresponding to various fault cause types, and combine the transmission line corridor monitoring information. Make corrections and give diagnostic results. The invention can accurately locate the fault of the transmission line and carry out the analysis of the cause of the tripping fault of the line, which can greatly reduce the workload of line inspection and improve the reliability of power supply.

The present invention is not limited to the specific embodiments described above, and the above is only a preferred embodiment of the present invention, and is not intended to limit the present invention. Any modifications, equivalents, etc. made within the spirit and principle of the present invention Substitutions and improvements, etc., should all be included within the protection scope of the present invention.

Claims (6)

1. An intelligent overhead transmission line fault type diagnosis method is characterized by comprising the following steps: the method comprises the following steps:

s1: judging whether a system with faults is provided with a distributed fault monitoring device and a power transmission line corridor environment monitoring device or not; if the distributed fault monitoring device and the power transmission line corridor environment monitoring device are installed, performing step S2, otherwise, performing step S3;

s2: searching the associated transient traveling wave characteristics to form a characteristic vector combination, inputting the characteristic vector combination into a first support vector machine, and carrying out lightning stroke and non-lightning stroke fault type diagnosis;

s21: if the lightning stroke fault is the lightning stroke fault, inquiring a lightning positioning system, combining fault positioning to obtain lightning current data closest to the fault tripping moment, and determining the lightning stroke fault probability P _lightning Correcting with lightning stroke fault probability correction function of f ₁ (ii) a For lightning stroke fault probability P _lightning The correction is specifically as follows:

P _{lightning-new} ＝f ₁ (P _{lightning-old} )＝a+(1-a)(P _{lightning-old} )；

wherein, P _{lightning-new} Indicating the corrected lightning stroke fault probability, P _{lightning-old} Representing the lightning stroke fault probability before correction, wherein a is a lightning stroke fault probability correction parameter;

the correction parameter a is comprehensively judged according to the amplitude A and the distance D of the lightning current nearest to the fault point, which are found in the lightning positioning system, and the calculation formula is as follows:

and correcting the fault probability of other reasons except the lightning stroke by adopting the corrected lightning stroke fault probability in the first vector machine, wherein the method specifically comprises the following steps:

wherein, P _other-new Indicating the probability of failure due to other causes, P, after correction _other-old Representing the fault probability of other reasons before correction;

s22: if the fault is a non-lightning fault, the system further queries an external online monitoring service platform comprising a forest fire monitoring system, an icing system and a meteorological system, calculates and extracts the power frequency waveform at the fault moment to obtain a characteristic combination, and inputs the characteristic combination into a second support vector machine to obtain the fault reason and the probability thereof;

correcting a fault result by combining mountain fire, windage yaw and icing data in the multi-source information;

s3: the system which has a fault is not provided with a distributed fault monitoring device and a power transmission line corridor environment monitoring device, and a characteristic combination which can be input into the Bayesian network is obtained by combining meteorological information according to the fault time, reclosing state and fault phase provided by the scheduling system information; and directly obtaining corresponding fault external feature prior probability distribution for the collected evidence information.

2. The intelligent overhead transmission line fault type diagnosis method according to claim 1, characterized in that: in step S22, the mountain fire in the multi-source information is used to correct the fault result, which specifically includes:

P _fire-new ＝f ₂ (P _fire-old )＝b+(1-b)(P _fire-old )；

wherein, P _fire-new Indicates the probability of mountain fire fault after correction, P _fire-old Indicates the probability of mountain fire failure before correction, f ₂ B is a mountain fire fault probability correction function, and b is a mountain fire fault probability correction parameter;

the calculation method of the mountain fire fault probability correction parameter b is as follows:

t is the fire point temperature.

3. The intelligent overhead transmission line fault type diagnosis method according to claim 1, characterized in that: in step S22, the windage yaw in the multi-source information is used to correct the fault result, which is specifically as follows:

P _wind-new ＝f ₃ (P _wind-old )＝c+(1-c)(P _wind-old )；

wherein, P _wind-new Indicating corrected windage yaw fault probability, P _wind-old Indicating windage yaw fault probability before correction, f ₃ C is a windage yaw fault probability correction function, and c is a windage yaw fault probability correction parameter;

the windage yaw fault probability correction parameter c is calculated as follows:

c＝[min(W,12)+2]/14；

w is the line corridor maximum wind speed.

4. The intelligent overhead transmission line fault type diagnosis method according to claim 1, characterized in that: in step S22, the icing in the multi-source information is used to correct the fault result, which is specifically as follows:

P _ice-new ＝f ₄ (P _ice-old )＝d+(1-d)(P _ice-old )；

wherein, P _ice-new Indicating the corrected icing fault probability, P _ice-old Representing the icing fault probability before correction, f ₄ D is an icing fault probability correction function and an icing fault probability correction parameter;

the icing fault probability correction parameter d is calculated as follows:

d＝0.5log(H+1)；

h is the maximum ice coating thickness monitored on-line.

5. The intelligent overhead transmission line fault type diagnosis method according to claim 1, characterized in that: the probability distribution in step S3 includes:

weather of failureIs event A; reclosing action is taken as an event B; the fault phase is event C; the failed month is event D; the failure time is event E; the fault wind power level is event F; assuming that the above events are independent of each other, the fault type is event V _i 1,2,3,4,5, which respectively represent a lightning stroke fault, a windage yaw fault, a bird damage fault, a tree flash fault and a mountain fire fault; the failure probability is calculated as follows:

and after the fault probability is obtained, inquiring the multi-source information again, performing the same correction on the fault probability by using the correction function, and outputting the final result.

6. The utility model provides an intelligence overhead transmission line fault type diagnostic system which characterized in that: the system comprises a system judgment module, a data module and a correction module; the system judgment module, the data module and the correction module are connected in sequence;

the system judgment module is used for judging whether a system with faults is provided with a distributed fault monitoring device and a power transmission line corridor environment monitoring device or not;

the data module is used for processing data according to the judgment result of the system judgment module, and specifically comprises the following steps:

if the system judgment module judges that a system with faults is provided with the distributed fault monitoring device and the transmission line corridor environment monitoring device, the data module searches the associated transient traveling wave characteristics to form a characteristic vector combination, and inputs the characteristic vector combination into a first vector machine to carry out lightning stroke and non-lightning stroke fault type diagnosis;

if the lightning stroke fault is diagnosed, the data module inquires a lightning positioning system, lightning current data closest to the fault tripping moment is obtained by combining fault positioning, and the correction module corrects the lightning stroke fault probability through a correction function;

if the non-lightning fault is diagnosed, the data module further queries an external online monitoring service platform comprising a forest fire monitoring system, an icing system and a meteorological system; calculating and extracting characteristics of the power frequency waveform at the fault moment to obtain a characteristic combination, and inputting the characteristic combination into a second support vector machine to obtain a fault reason and the probability thereof; the correction module corrects the fault result by combining mountain fire, windage yaw and icing data in the multi-source information;

if the system judgment module judges that a system with a fault is not provided with the distributed fault monitoring device and the transmission line corridor environment monitoring device, the data module combines meteorological information according to the fault time, the reclosing state and the fault phase provided by the scheduling system information to obtain a characteristic combination which can be input into the Bayesian network; directly obtaining corresponding fault external feature prior probability distribution for collected evidence information;

the correction function of the correction module for the lightning stroke fault probability is as follows:

P _{lightning-new} ＝f ₁ (P _{lightning-old} )＝a+(1-a)(P _{lightning-old} )；

wherein, P _{lightning-new} Indicating the corrected lightning stroke fault probability, P _{lightning-old} Representing the lightning stroke fault probability before correction, wherein a is a lightning stroke fault probability correction parameter; f. of ₁ Correcting a function for lightning stroke fault probability; the correction parameter a is comprehensively judged according to the amplitude A and the distance D of the lightning current nearest to the fault point, which are detected in the lightning positioning system, and the calculation formula is as follows:

Images