CN104573866B - Method and system for predicting defects of electric power equipment - Google Patents

Abstract

The present invention provides a method for predicting the defect of electric equipment, which is implemented on N similar electric equipment, including: extracting historical data of electric equipment in a period of time, and obtaining K defect analysis variables satisfying predetermined conditions; The duration under the same defect analysis variable and the total duration under K defect analysis variables are further obtained to obtain the initial probability of each defect state variable and combined into a vector P(0); to obtain the total number of K defect analysis variables, And in the evenly divided into M time slices, determine the conversion times of K*K kinds of defect analysis variables between adjacent time slices, and further obtain K*K conversion probabilities and combine them into a matrix P; according to P(1)=P(0 )*P, determine the defect category corresponding to the maximum value in P(1) as the predicted defect before the new power equipment is put into use. The implementation of the present invention can accurately predict the possible defects of equipment, reduce the labor load of equipment operation and maintenance personnel, and improve the actual guidance effect of defect analysis.

Description

A method and system for predicting defects in electrical equipment

technical field

The invention relates to the technical field of electric equipment management, in particular to a method and system for predicting electric equipment defects.

Background technique

Healthy power equipment is the basis for ensuring the normal operation of the grid and improving the reliability of the grid power supply. In actual operation, although some power equipment can continue to be used, if its operation status is abnormal or there are hidden dangers, it will shorten the life of the equipment, affect the safety of people, equipment and the power grid, and lead to adverse situations such as degradation of power quality. In summary, the abnormalities or hidden dangers of the above-mentioned electrical equipment are collectively referred to as defects. Therefore, for all kinds of defects in power equipment, it is necessary to find and eliminate them in time to prevent defects from developing into faults and causing bad situations such as power grid outages.

At present, the analysis methods for defects in power equipment mainly include: statistics on the number of defects, defect elimination rate and timely rate of defect elimination in electric The overall defect rate of equipment is predicted. Although this analysis and prediction method can grasp the occurrence and development of equipment defects to a certain extent, the disadvantage is that the information is not accurate enough to effectively guide the development of equipment inspection work. For example, the prediction result of the overall defect rate is the proportion of the number of devices that may have defects in a certain batch of electrical equipment in the future to the total number, but in actual production, it cannot effectively guide the work of equipment operation and maintenance personnel.

Contents of the invention

The technical problem to be solved by the embodiments of the present invention is to provide a method and system for predicting defects of electric equipment, which can accurately predict the possible defects of electric equipment, and help the operation and maintenance personnel of electric equipment carry out focused inspection and maintenance in a targeted manner. It reduces the workload of power equipment operation and maintenance personnel, and improves the guidance effect of power equipment defect analysis on actual production work.

In order to solve the above technical problems, an embodiment of the present invention provides a method for predicting defects of electrical equipment, which is implemented on N electrical equipment of the same type, and the method includes:

Step a, extracting the historical data of the N electric equipments in a period of time, obtaining K defect categories satisfying predetermined conditions according to the statistics of the historical data, and using the obtained K defect categories as defect analysis variables; Wherein, K and N are both positive integers;

Step b. In the historical data, for each defect analysis variable, accumulate the duration of the N electric equipment under the same defect analysis variable, and accumulate K pieces of the N electric equipment under the same defect analysis variable The total duration is obtained by the following duration, and the initial probability of each defect state variable is determined according to the accumulated total duration and the accumulated duration corresponding to each defect analysis variable, and further determined by the K initial probabilities are combined into a vector P(0);

Step c. Obtain the total number of occurrences of the K defect analysis variables respectively in the historical data, divide the time period into M time slices, and sequentially determine the N power values between adjacent time slices The equipment corresponds to the number of conversion modes of K*K defect analysis variables, and according to the total number of appearances of K defect analysis variables acquired and the number of conversion modes of K*K defect analysis variables, determine the K*K types The conversion probabilities corresponding to the conversion modes of the defect analysis variables are further combined into a matrix P from the K*K conversion probabilities; wherein, M is a positive integer;

Step d. Determine the maximum value in P(1) according to the formula P(1)=P(0)*P, and use the defect category corresponding to the determined maximum value as the predicted defect before the new electric equipment is put into use.

Wherein, the method further includes:

Obtain the defects currently occurring in a power equipment that has been put into use, and find the maximum value corresponding to the K conversion methods in the matrix P according to the type of the current defect, and assign the maximum value corresponding to the found value to The defect category is used as the next predicted defect of the current defect-occurring electrical equipment.

Wherein, the concrete steps of described step a include:

Extracting the historical data of the N electrical equipment in a time period, and determining the total number of occurrences of each defect in the historical data;

Arrange the total number of occurrences of each defect in order from large to small, screen out the first K-1 defects with the largest total number of times, and use the category corresponding to the K-1 defects and the category without defects as the ones that meet the conditions K defect categories are further set as defect analysis variables.

Wherein, the concrete steps of described step b include:

In the historical data, the continuous occurrence time corresponding to each defect analysis variable of the N electric equipment is obtained, and the duration of the N electric equipment under the same defect analysis variable is screened out and accumulated to obtain K defect analysis The durations corresponding to the variables respectively;

Accumulate the duration corresponding to the obtained K defect analysis variables respectively to obtain the total duration;

Dividing the durations corresponding to the obtained K defect analysis variables respectively by the obtained total duration to obtain initial probabilities corresponding to the K defect analysis variables respectively, and combining the K initial probabilities into a vector P(0).

Wherein, the concrete steps of described step c include:

Obtaining the total number of occurrences of the K defect analysis variables respectively in the historical data;

Perform pairwise mapping on the K defect analysis variables to obtain K*K defect analysis variable conversion modes;

Dividing the time period into M time slices, sorting the time from small to large, and determining the number of conversion times of the N electrical equipment corresponding to K*K defect analysis variable conversion modes between adjacent time slices;

The number of conversions of each defect analysis variable is used as the molecule, and the defect analysis variable corresponding to the main map in each molecule is determined, and the total number of occurrences of the defect analysis variable corresponding to the main map in each molecule is selected as Corresponding to the denominator, the conversion probabilities corresponding to the K*K defect analysis variable conversion modes are obtained, and the K*K conversion probabilities are further combined into a matrix P.

Wherein, the electric equipment is a transformer, and the defect analysis variables include no defect, oil leakage, cooling system failure, instrument failure, operating mechanism abnormality and external mechanical damage.

The embodiment of the present invention further provides a method for predicting defects of electrical equipment, which is implemented on N electrical equipment of the same type, and the method includes:

Step S1, extracting the historical data of the N electrical equipment in a period of time, and obtaining K defect categories satisfying predetermined conditions according to the statistics of the historical data, and using the obtained K defect categories as defect analysis variables; Wherein, K and N are both positive integers;

Step S2. Obtain the total number of occurrences of the K defect analysis variables respectively in the historical data, divide the time period into M time slices, and sequentially determine the N power values between adjacent time slices The equipment corresponds to the number of conversion modes of K*K defect analysis variables, and according to the total number of appearances of K defect analysis variables acquired and the number of conversion modes of K*K defect analysis variables, determine the K*K types The conversion probabilities corresponding to the conversion modes of the defect analysis variables are further combined into a matrix P from the K*K conversion probabilities; wherein, M is a positive integer;

Step S3. Obtain the current defect of a power equipment that has been put into use, and find the maximum value corresponding to the K conversion methods in the matrix P according to the type of the current defect, and convert the found maximum value to The defect category corresponding to the value is used as the next predicted defect of the electric equipment that currently has a defect.

Wherein, the concrete steps of described step S1 include:

Extracting the historical data of the N electrical equipment in a time period, and determining the total number of occurrences of each defect in the historical data;

Arrange the total number of occurrences of each defect in order from large to small, screen out the first K-1 defects with the largest total number of times, and use the category corresponding to the K-1 defects and the category without defects as the ones that meet the conditions K defect categories are further set as defect analysis variables.

Wherein, the concrete steps of described step S2 include:

Obtaining the total number of occurrences of the K defect analysis variables respectively in the historical data;

Perform pairwise mapping on the K defect analysis variables to obtain K*K defect analysis variable conversion modes;

Dividing the time period into M time slices, sorting the time from small to large, and determining the number of conversion times of the N electrical equipment corresponding to K*K defect analysis variable conversion modes between adjacent time slices;

The number of conversions of each defect analysis variable is used as the molecule, and the defect analysis variable corresponding to the main map in each molecule is determined, and the total number of occurrences of the defect analysis variable corresponding to the main map in each molecule is selected as Corresponding to the denominator, the conversion probabilities corresponding to the K*K defect analysis variable conversion modes are obtained, and the K*K conversion probabilities are further combined into a matrix P.

Wherein, the method further includes:

In the historical data, for each defect analysis variable, the duration of the N electric equipment under the same defect analysis variable is accumulated, and K durations of the N electric equipment under the same defect analysis variable are accumulated time to obtain the total duration, and according to the accumulated total duration and the accumulated duration corresponding to each defect analysis variable, determine the initial probability of each defect state variable, and further from the K initial The probabilities are combined into a vector P(0);

According to the formula P(1)=P(0)*P, the maximum value in P(1) is determined, and the defect category corresponding to the determined maximum value is used as the predicted defect before the new electric equipment is put into use.

Wherein, in the historical data, for each defect analysis variable, the duration of the N electric equipment under the same defect analysis variable is accumulated, and K number of the N electric equipment under the same defect analysis variable are accumulated. The total duration is obtained by the duration under the variable, and the initial probability of each defect state variable is determined according to the accumulated total duration and the accumulated duration corresponding to each defect analysis variable, and further obtained from the The specific steps for combining the K initial probabilities into vector P(0) include:

In the historical data, the continuous occurrence time corresponding to each defect analysis variable of the N electric equipment is obtained, and the duration of the N electric equipment under the same defect analysis variable is screened out and accumulated to obtain K defect analysis The durations corresponding to the variables respectively;

Accumulate the duration corresponding to the obtained K defect analysis variables respectively to obtain the total duration;

Dividing the durations corresponding to the obtained K defect analysis variables respectively by the obtained total duration to obtain initial probabilities corresponding to the K defect analysis variables respectively, and combining the K initial probabilities into a vector P(0).

The embodiment of the present invention also provides a system for predicting defects of electrical equipment, which is implemented on N electrical equipment of the same type, and the system includes:

Determining a defect analysis variable unit for extracting the historical data of the N electrical equipment in a period of time, obtaining K defect categories satisfying predetermined conditions according to the statistics of the historical data, and dividing the obtained K defect categories As a defect analysis variable; among them, K and N are both positive integers;

Acquiring an initial probability unit for accumulating the duration of the N electric equipment under the same defect analysis variable for each defect analysis variable in the historical data, and accumulating K pieces of the N electric equipment at The total duration is obtained from the duration under the same defect analysis variable, and the initial probability of each defect state variable is determined according to the accumulated total duration and the accumulated duration corresponding to each defect analysis variable, Further be combined into vector P (0) by described K initial probabilities;

Obtaining a conversion probability unit for obtaining the total number of occurrences of the K defect analysis variables in the historical data, and dividing the time period into M time slices, and sequentially determining the number of times between adjacent time slices The N electric equipments correspond to the number of times of conversion of K*K kinds of defect analysis variables, and according to the total number of occurrences of the acquired K defect analysis variables and the number of K*K kinds of defect analysis variable conversion times, determine the The conversion probabilities corresponding to K*K kinds of defect analysis variable conversion modes are further combined into a matrix P by the K*K conversion probabilities; wherein, M is a positive integer;

The new equipment prediction defect unit is used to determine the maximum value in P(1) according to the formula P(1)=P(0)*P, and put the defect category corresponding to the determined maximum value into use as new electric equipment previous prediction deficiencies.

Wherein, the system also includes:

The operating equipment prediction defect unit is used to obtain the current defects of a power device that has been put into use, and find the maximum value corresponding to the K conversion methods in the matrix P according to the type of the current defect, and convert the The defect category corresponding to the found maximum value is used as the next predicted defect of the electric equipment that currently has a defect.

The embodiment of the present invention further provides a system for predicting defects of electrical equipment, which is implemented on N electrical equipment of the same type, and the system includes:

Determining a defect analysis variable unit for extracting the historical data of the N electrical equipment in a period of time, obtaining K defect categories satisfying predetermined conditions according to the statistics of the historical data, and dividing the obtained K defect categories As a defect analysis variable; among them, K and N are both positive integers;

Obtaining a conversion probability unit for obtaining the total number of occurrences of the K defect analysis variables in the historical data, and dividing the time period into M time slices, and sequentially determining the number of times between adjacent time slices The N electric equipments correspond to the number of times of conversion of K*K kinds of defect analysis variables, and according to the total number of occurrences of the acquired K defect analysis variables and the number of K*K kinds of defect analysis variable conversion times, determine the The conversion probabilities corresponding to K*K kinds of defect analysis variable conversion modes are further combined into a matrix P by the K*K conversion probabilities; wherein, M is a positive integer;

The operating equipment prediction defect unit is used to obtain the current defects of a power device that has been put into use, and find the maximum value corresponding to the K conversion methods in the matrix P according to the type of the current defect, and convert the The defect category corresponding to the found maximum value is used as the next predicted defect of the electric equipment that currently has a defect.

Wherein, the system also includes:

Acquiring an initial probability unit for accumulating the duration of the N electric equipment under the same defect analysis variable for each defect analysis variable in the historical data, and accumulating K pieces of the N electric equipment at The total duration is obtained from the duration under the same defect analysis variable, and the initial probability of each defect state variable is determined according to the accumulated total duration and the accumulated duration corresponding to each defect analysis variable, Further be combined into vector P (0) by described K initial probabilities;

The new equipment prediction defect unit is used to determine the maximum value in P(1) according to the formula P(1)=P(0)*P, and put the defect category corresponding to the determined maximum value into use as new electric equipment previous prediction deficiencies.

Implementing the embodiment of the present invention has the following beneficial effects:

In the embodiment of the present invention, based on the historical defect information and current operating conditions of similar electric equipment, the Markov prediction algorithm is used to accurately predict the possible defects of electric equipment (including those that have not been put into use and those that have been put into use and have had defects) , can help power equipment operation and maintenance personnel to conduct targeted inspections and maintenance, reduce the workload of power equipment operation and maintenance personnel, and improve the guidance effect of power equipment defect analysis on actual production work.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the 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. For those of ordinary skill in the art, obtaining other drawings based on these drawings still belongs to the scope of the present invention without any creative effort.

FIG. 1 is a flow chart of a method for predicting defects in electrical equipment provided by an embodiment of the present invention;

FIG. 2 is a flow chart of another method for predicting defects in electrical equipment provided by an embodiment of the present invention;

FIG. 3 is a schematic structural diagram of a system for predicting defects of electrical equipment provided by an embodiment of the present invention;

Fig. 4 is a schematic structural diagram of another system for predicting defects of electrical equipment provided by an embodiment of the present invention.

detailed description

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings.

As shown in FIG. 1 , an embodiment of the present invention provides a method for predicting defects of electrical equipment, which is implemented on N electrical equipment of the same type, and the method includes:

Step S101, extracting the historical data of the N electrical equipment in a period of time, and obtaining K defect categories satisfying predetermined conditions according to the statistics of the historical data, and using the obtained K defect categories as defect analysis variables; Wherein, K and N are both positive integers;

The specific process is to extract the historical data of N electrical equipment in a period of time, and determine the total number of occurrences of each defect in the historical data;

Arrange the total number of occurrences of each defect in descending order, screen out the first K-1 defects with the largest total number of times, and use the category corresponding to the K-1 defects and the category without defects as the K that meet the conditions The defect category is further set as a defect analysis variable.

As an example, the power equipment is a transformer, and the main reasons for its defects include but are not limited to oil leakage, cooling system failure, instrument failure, abnormal operating mechanism, external mechanical damage, etc. Therefore, when extracting the operating data of multiple transformers within one year , the operating data includes defect data with the above-mentioned defects. Count the total number of occurrences of all defects in all transformers in the historical data, and then sort all the defects in descending order of the total number of times, and filter out the top defects as oil leakage, cooling system failure, instrumentation Failure, operating mechanism abnormality and external mechanical damage, and the screened defect categories and non-defect categories are used as the corresponding defect analysis variables. The sequence of defect analysis variables is defined as no defect, oil leakage, cooling system failure, Instrument failure, abnormal operating mechanism and external mechanical damage.

Step S102, in the historical data, for each defect analysis variable, accumulate the duration of the N electric equipment under the same defect analysis variable, and accumulate K pieces of the N electric equipment under the same defect analysis variable The total duration is obtained by the following duration, and the initial probability of each defect state variable is determined according to the accumulated total duration and the accumulated duration corresponding to each defect analysis variable, and further determined by the K initial probabilities are combined into a vector P(0);

The specific process is, in the historical data, obtain the continuous occurrence time of N electric equipment corresponding to each defect analysis variable, screen out and accumulate the duration of N electric equipment under the same defect analysis variable, and obtain K defect analysis variables respectively the corresponding duration;

Accumulate the corresponding durations of the K defect analysis variables respectively to obtain the total duration;

Divide the durations corresponding to the obtained K defect analysis variables by the obtained total duration to obtain the initial probabilities corresponding to the K defect analysis variables, and combine the K initial probabilities into a vector P(0).

In the embodiment of the present invention, the K initial probabilities can be expressed as p ₁ (0), p ₂ (0), ...p _K (0), so the vector P(0)=(p ₁ (0), p ₂ (0), . . . p _K (0)).

Taking the aforementioned transformer as an example, p ₁ (0) represents the probability of no defect, p ₂ (0), ...p ₆ (0) represent oil leakage, cooling system failure, instrument failure, operating mechanism abnormality and external Initial probability of defects such as mechanical damage. The steps to calculate the initial probability of the defect analysis variables are as follows (taking p ₁ (0) as an example):

Accumulate the duration of all transformers without defects, accumulate the duration of all transformer oil leakage, accumulate the duration of all transformer cooling system fault defects, accumulate the duration of all transformer instrument failures, and accumulate the duration of all transformer operating mechanism abnormalities, Until the result of accumulating the duration of external mechanical damage of all transformers is obtained;

The total duration is obtained by summing up the accumulated duration of the above six defect analysis variables;

Taking the duration of all transformers without defects as the numerator and the total duration as the denominator, p ₁ (0) is obtained.

Similarly, p ₂ (0), ... p ₆ (0) can be obtained according to the above method, so as to further form the vector P (0) = (p ₁ (0), p ₂ (0), ... p ₆ (0)).

Step S103: Obtain the total number of occurrences of the K defect analysis variables respectively in the historical data, divide the time period into M time slices, and sequentially determine the N power values between adjacent time slices The equipment corresponds to the number of conversion modes of K*K defect analysis variables, and according to the total number of appearances of K defect analysis variables acquired and the number of conversion modes of K*K defect analysis variables, determine the K*K types The conversion probabilities corresponding to the conversion modes of the defect analysis variables are further combined into a matrix P from the K*K conversion probabilities; wherein, M is a positive integer;

The specific process is to obtain the total number of occurrences of the K defect analysis variables respectively in the historical data;

Perform pairwise mapping on K defect analysis variables to obtain K*K defect analysis variable conversion methods;

Divide the time period into M time slices, sort them from small to large, and determine the number of times that N electrical equipment between adjacent time slices corresponds to K*K kinds of defect analysis variable conversion methods;

The number of conversions of each defect analysis variable is taken as the numerator, and the defect analysis variable corresponding to the main map in each molecule is determined, and the total number of occurrences of the defect analysis variable corresponding to the main map in each molecule is selected as the corresponding The denominator is used to obtain the conversion probabilities corresponding to the K*K defect analysis variable conversion modes, and further combine the K*K conversion probabilities into a matrix P.

It should be noted that, for calculation purposes, K defect analysis variables are pairwise mapped, which includes the mutual mapping of the same defect analysis variable, such as the leakage oil variable of a transformer is mapped to the same leakage oil variable, so the formed matrix .

It should be noted that the M time slices that divide the time period evenly are as small as possible, so that the number of defect category transitions in adjacent time slices (from the previous time slice to the next time slice) satisfies one time, such as the last time slice A The transformer has oil leakage, but in the next time slice A transformer will only have one cooling system failure defect, not more than one cooling system failure defect.

In the embodiment of the present invention, the conversion probabilities corresponding to K*K defect analysis variable conversion modes can be expressed as (p ₁₁ , p ₁₂ , ... p _1K ; p ₂₁ , p ₂₂ , ... p _2K ; ...p _K1 , p _K2 , ...p _KK ), thus forming the matrix Wherein, p _nn represents that the defect categories corresponding to the defect analysis variables of the same electric equipment in adjacent time slices are the same, n=1,2,...K.

Taking the aforementioned transformer as an example, a year is divided into twelve months to obtain twelve time slices; p ₁₁ indicates the conversion probability that the equipment defect status between the first and last two months is no defect, and p ₁₂ indicates the transition probability of the previous two months The conversion probability of no defect in one time slice and the occurrence of oil leakage defect in the next time slice, so by analogy, the conversion probabilities corresponding to the conversion modes of the transformer defect analysis variables can be expressed as (p ₁₁ , p ₁₂ , . ..p ₁₆ ; p ₂₁ , p ₂₂ , . . . p ₂₆ ; . . . p ₆₁ , p ₆₂ , . . . p ₆₆ ). The steps for calculating the transition probability of the defect analysis variables are as follows (taking no defects (p ₁₁ , p ₁₂ , ... p ₁₆ ) as an example):

Obtain the total number of occurrences of the six defect analysis variables in the historical data. Of course, for (p ₁₁ , p ₁₂ , ... p ₁₆ ), it is only necessary to determine that no defects occur in all transformers in the historical data total number of times;

No defect will be mapped to six conversion methods: no defect mapping no defect, no defect mapping oil leakage, no defect mapping cooling system failure, no defect mapping instrument failure, no defect mapping operating mechanism Abnormal and non-occurring defects map external mechanical damage;

Sort from January to December, and determine the number of conversion times of all transformers corresponding to the above six defect analysis variables between adjacent time slices;

The times of the above-mentioned six defect analysis variable conversion modes are taken as the numerator, and the total times of all transformers without defects in the historical data are taken as the denominator, and (p ₁₁ , p ₁₂ ,...p ₁₆ ) are obtained.

Similarly, (p ₂₁ , p ₂₂ , ... p ₂₆ ; ... p ₆₁ , p ₆₂ , ... p ₆₆ ) can be obtained according to the above method, so as to further form the matrix

Step S104, according to the formula P(1)=P(0)*P, determine the maximum value in P(1), and use the defect category corresponding to the determined maximum value as the predicted defect before the new electric equipment is put into use.

By predicting the defects of the new power equipment before it is put into use, the maintenance personnel of the power equipment can carry out targeted inspection and maintenance, reduce the labor of the operation and maintenance personnel of the power equipment, and improve the guidance effect of the defect analysis of the power equipment on the actual production work.

Taking the aforementioned transformer as an example, It is obtained that the maximum value in P(1) corresponds to the fault defect of the cooling system, and the fault defect of the cooling system is regarded as the defect of the predicted transformer.

The transformation matrix generated by the method for predicting the defects of electric equipment in the embodiment of the present invention also has a guiding effect on the actual production work for the defect prediction of the electric equipment that has been put into use, and under certain conditions, it does not need to increase the calculation amount of data for the current Electric equipment performs defect prediction, so the method further includes:

Obtain the currently occurring defects of a power equipment that has been put into use, and find the maximum value corresponding to the K conversion methods in the matrix P according to the type of the current defect, and use the defect type corresponding to the found maximum value as the current occurrence The next predicted defect for a defective electrical device.

As shown in FIG. 2, the embodiment of the present invention provides another method for predicting defects of electrical equipment, which is implemented on N electrical equipment of the same type, and the method includes:

Step S201, extracting the historical data of the N electrical equipment in a period of time, and obtaining K defect categories satisfying predetermined conditions according to the statistics of the historical data, and using the obtained K defect categories as defect analysis variables; Wherein, K and N are both positive integers;

The specific process is to extract the historical data of N electrical equipment in a period of time, and determine the total number of occurrences of each defect in the historical data;

Arrange the total number of occurrences of each defect in descending order, screen out the first K-1 defects with the largest total number of times, and use the category corresponding to the K-1 defects and the category without defects as the K that meet the conditions The defect category is further set as a defect analysis variable.

Step S202. Obtain the total number of occurrences of the K defect analysis variables respectively in the historical data, divide the time period into M time slices, and sequentially determine the N power values between adjacent time slices. The equipment corresponds to the number of conversion modes of K*K defect analysis variables, and according to the total number of appearances of K defect analysis variables acquired and the number of conversion modes of K*K defect analysis variables, determine the K*K types The conversion probabilities corresponding to the conversion modes of the defect analysis variables are further combined into a matrix P from the K*K conversion probabilities; wherein, M is a positive integer;

The specific process is to obtain the total number of occurrences of the K defect analysis variables respectively in the historical data;

Perform pairwise mapping on K defect analysis variables to obtain K*K defect analysis variable conversion methods;

Divide the time period into M time slices, sort them from small to large, and determine the number of times that N electrical equipment between adjacent time slices corresponds to K*K kinds of defect analysis variable conversion methods;

The number of conversions of each defect analysis variable is taken as the numerator, and the defect analysis variable corresponding to the main map in each molecule is determined, and the total number of occurrences of the defect analysis variable corresponding to the main map in each molecule is selected as the corresponding The denominator is used to obtain the conversion probabilities corresponding to the K*K defect analysis variable conversion modes, and further combine the K*K conversion probabilities into a matrix P.

It should be noted that, for calculation purposes, K defect analysis variables are pairwise mapped, which includes the mutual mapping of the same defect analysis variable, such as the leakage oil variable of a transformer is mapped to the same leakage oil variable, so the formed matrix .

It should be noted that the M time slices that divide the time period evenly are as small as possible, so that the number of defect category transitions in adjacent time slices (from the previous time slice to the next time slice) satisfies one time, such as the last time slice A The transformer has oil leakage, but in the next time slice A transformer will only have one cooling system failure defect, not more than one cooling system failure defect.

In the embodiment of the present invention, the conversion probabilities corresponding to K*K defect analysis variable conversion modes can be expressed as (p ₁₁ , p ₁₂ , ... p _1K ; p ₂₁ , p ₂₂ , ... p _2K ; ...p _K1 , p _K2 , ...p _KK ), thus forming the matrix Wherein, p _nn represents that the defect categories corresponding to the defect analysis variables of the same electric equipment in adjacent time slices are the same, n=1,2,...K.

Step S203: Obtain the current defect of a power equipment that has been put into use, and find the maximum value corresponding to the K conversion methods in the matrix P according to the type of the current defect, and convert the found maximum value to The defect category corresponding to the value is used as the next predicted defect of the electric equipment that currently has a defect.

By predicting the defects of the power equipment that has been put into use before it is put into use, the maintenance personnel of the power equipment can make targeted inspections and maintenance, reduce the workload of the operation and maintenance personnel of the power equipment, and improve the guidance of the defect analysis of the power equipment for the actual production work Effect.

In the embodiment of the present invention, predicting the defects of new electric equipment before putting into use also has a guiding effect on actual production work, so the method further includes:

Step S21, in the historical data, for each defect analysis variable, accumulate the duration of the N electric equipment under the same defect analysis variable, and accumulate K pieces of the N electric equipment under the same defect analysis variable The total duration is obtained by the following duration, and the initial probability of each defect state variable is determined according to the accumulated total duration and the accumulated duration corresponding to each defect analysis variable, and further determined by the K initial probabilities are combined into a vector P(0);

The specific process is, in the historical data, obtain the continuous occurrence time of N electric equipment corresponding to each defect analysis variable, screen out and accumulate the duration of N electric equipment under the same defect analysis variable, and obtain K defect analysis variables respectively the corresponding duration;

Accumulate the corresponding durations of the K defect analysis variables respectively to obtain the total duration;

Divide the durations corresponding to the obtained K defect analysis variables by the obtained total duration to obtain the initial probabilities corresponding to the K defect analysis variables, and combine the K initial probabilities into a vector P(0).

In the embodiment of the present invention, the K initial probabilities can be expressed as p ₁ (0), p ₂ (0), ...p _K (0), so the vector P(0)=(p ₁ (0), p ₂ (0), . . . p _K (0)).

Step S22, according to the formula P(1)=P(0)*P, determine the maximum value in P(1), and use the defect category corresponding to the determined maximum value as the predicted defect before the new electric equipment is put into use.

Wherein, in the historical data, for each defect analysis variable, the duration of the N electric equipment under the same defect analysis variable is accumulated, and K number of the N electric equipment under the same defect analysis variable are accumulated. The total duration is obtained by the duration under the variable, and the initial probability of each defect state variable is determined according to the accumulated total duration and the accumulated duration corresponding to each defect analysis variable, and further obtained from the The specific steps for combining the K initial probabilities into vector P(0) include:

As shown in Figure 3, the embodiment of the present invention also provides a system for predicting defects of electric equipment, which is implemented on N electric equipment of the same type, and the system includes:

Determining the defect analysis variable unit 110, configured to extract the historical data of the N electrical equipment in a period of time, obtain K defect categories satisfying predetermined conditions according to the statistics of the historical data, and convert the obtained K defect categories Both are used as defect analysis variables; among them, K and N are both positive integers;

Acquire the initial probability unit 120, which is used to accumulate the duration of the N electric equipment under the same defect analysis variable for each defect analysis variable in the historical data, and accumulate K pieces of the N electric equipment The total duration is obtained from the duration under the same defect analysis variable, and the initial probability of each defect state variable is determined according to the accumulated total duration and the accumulated duration corresponding to each defect analysis variable , further combining the K initial probabilities into a vector P(0);

Obtaining the transition probability unit 130, configured to obtain the total number of occurrences of the K defect analysis variables in the historical data, and divide the time period into M time slices, and sequentially determine the time interval between adjacent time slices. The N electrical equipments correspond to the number of conversion modes of K*K defect analysis variables, and according to the total number of occurrences of the acquired K defect analysis variables and the number of conversion modes of K*K defect analysis variables, determine the The conversion probabilities corresponding to the K*K defect analysis variable conversion modes are further combined into a matrix P by the K*K conversion probabilities; wherein, M is a positive integer;

The new equipment prediction defect unit 140 is used to determine the maximum value in P(1) according to the formula P(1)=P(0)*P, and use the defect category corresponding to the determined maximum value as new electric equipment input Predict defects before use.

Among them, the system also includes:

The operating equipment prediction defect unit 150 is used to obtain the current defect of a power device that has been put into use, and find the maximum value corresponding to the K conversion methods in the matrix P according to the type of the current defect, and The defect category corresponding to the found maximum value is used as the next predicted defect of the electric equipment that currently has a defect.

Wherein, obtaining the initial probability unit 120 includes:

The obtaining duration module 1201 is configured to obtain, from the historical data, the duration of occurrence of the N electric equipment corresponding to each defect analysis variable, filter out the duration of the N electric equipment under the same defect analysis variable, and Accumulate to obtain the durations corresponding to the K defect analysis variables respectively;

Obtaining the total duration module 1202, configured to accumulate the obtained duration corresponding to the K defect analysis variables respectively, to obtain the total duration;

Determine the initial probability module 1203, which is used to divide the obtained durations corresponding to the K defect analysis variables respectively by the obtained total duration to obtain the initial probabilities respectively corresponding to the K defect analysis variables, and calculate the obtained The above K initial probabilities are combined into a vector P(0).

Wherein, obtaining conversion probability unit 130 includes:

Obtaining the total number of defects module 1301, configured to obtain the total number of occurrences of the K defect analysis variables respectively in the historical data;

Defect mapping module 1302, configured to perform pairwise mapping on the K defect analysis variables to obtain K*K defect analysis variable conversion modes;

Obtaining defect conversion times module 1303, which is used to divide the time period into M time slices, and sort them from small to large, and determine that the N electrical equipment between adjacent time slices corresponds to K*K kinds of defect analysis variables the number of conversions;

Determine the conversion probability module 1304, which is used to use the number of conversions of each defect analysis variable as a numerator, and determine the defect analysis variable corresponding to the main map in each molecule, and screen out the corresponding defect analysis variable in each molecule. The total number of occurrences of the defect analysis variables is used as the corresponding denominator to obtain the conversion probabilities corresponding to the K*K conversion modes of the defect analysis variables, and further combine the K*K conversion probabilities into a matrix P.

As shown in Figure 4, the embodiment of the present invention also provides another system for predicting defects of electric equipment, which is implemented on N electric equipment of the same type, and the system includes:

Determining the defect analysis variable unit 210, configured to extract the historical data of the N electrical equipment in a period of time, obtain K defect categories satisfying predetermined conditions according to the statistics of the historical data, and convert the obtained K defect categories Both are used as defect analysis variables; among them, K and N are both positive integers;

Obtaining conversion probability unit 220, configured to obtain the total number of occurrences of the K defect analysis variables respectively in the historical data, divide the time period into M time slices, and sequentially determine the time interval between adjacent time slices The N electrical equipments correspond to the number of conversion modes of K*K defect analysis variables, and according to the total number of occurrences of the acquired K defect analysis variables and the number of conversion modes of K*K defect analysis variables, determine the The conversion probabilities corresponding to the K*K defect analysis variable conversion modes are further combined into a matrix P by the K*K conversion probabilities; wherein, M is a positive integer;

The operating equipment prediction defect unit 230 is used to obtain the current defect of a power device that has been put into use, and find the maximum value corresponding to the K conversion methods in the matrix P according to the type of the current defect, and set The defect category corresponding to the found maximum value is used as the next predicted defect of the electric equipment currently having a defect.

Wherein, the acquisition conversion probability unit 220 includes:

Obtaining the total number of defects module 2201, configured to obtain the total number of occurrences of the K defect analysis variables respectively in the historical data;

Defect mapping module 2202, configured to perform pairwise mapping on the K defect analysis variables to obtain K*K defect analysis variable conversion modes;

Obtaining defect conversion times module 2203, which is used to divide the time period into M time slices, and sort them from small to large, and determine that the N electrical equipment between adjacent time slices corresponds to K*K kinds of defect analysis variables the number of conversions;

Determine the conversion probability module 2204, which is used to use the number of conversions of each defect analysis variable as a numerator, and determine the defect analysis variable corresponding to the main map in each molecule, and screen out the corresponding defect analysis variable in each molecule. The total number of occurrences of the defect analysis variables is used as the corresponding denominator to obtain the conversion probabilities corresponding to the K*K conversion modes of the defect analysis variables, and further combine the K*K conversion probabilities into a matrix P.

Wherein, the system also includes:

Acquiring the initial probability unit 240, which is used to accumulate the duration of the N electric equipment under the same defect analysis variable for each defect analysis variable in the historical data, and accumulate K pieces of the N electric equipment The total duration is obtained from the duration under the same defect analysis variable, and the initial probability of each defect state variable is determined according to the accumulated total duration and the accumulated duration corresponding to each defect analysis variable , further combining the K initial probabilities into a vector P(0);

The new equipment prediction defect unit 250 is used to determine the maximum value in P(1) according to the formula P(1)=P(0)*P, and use the defect category corresponding to the determined maximum value as new electric equipment input Predict defects before use.

Wherein, the obtaining initial probability unit 240 includes:

The obtaining duration module 2401 is configured to obtain, from the historical data, the duration of occurrence of the N electric equipment corresponding to each defect analysis variable, filter out the duration of the N electric equipment under the same defect analysis variable, and Accumulate to obtain the durations corresponding to the K defect analysis variables respectively;

Obtaining the total duration module 2402, configured to accumulate the durations corresponding to the obtained K defect analysis variables to obtain the total duration;

Determine the initial probability module 2403, which is used to divide the obtained durations corresponding to the K defect analysis variables respectively by the obtained total duration to obtain the initial probabilities respectively corresponding to the K defect analysis variables, and calculate the obtained The above K initial probabilities are combined into a vector P(0).

Implementing the embodiment of the present invention has the following beneficial effects:

In the embodiment of the present invention, based on the historical defect information and current operating conditions of similar electric equipment, the Markov prediction algorithm is used to accurately predict the possible defects of electric equipment (including those that have not been put into use and those that have been put into use and have had defects) , can help power equipment operation and maintenance personnel to conduct targeted inspections and maintenance, reduce the workload of power equipment operation and maintenance personnel, and improve the guidance effect of power equipment defect analysis on actual production work.

It is worth noting that in the above system embodiments, the system units included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific functions of each functional unit The names are only for the convenience of distinguishing each other, and are not used to limit the protection scope of the present invention.

Those of ordinary skill in the art can understand that all or part of the steps in the method of the above-mentioned embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, and the storage Media such as ROM/RAM, magnetic disk, optical disk, etc.

The above disclosures are only preferred embodiments of the present invention, and certainly cannot limit the scope of rights of the present invention. Therefore, equivalent changes made according to the claims of the present invention still fall within the scope of the present invention.

Claims (15)

1. A method for predicting electrical equipment defects, characterized in that it is implemented on N electrical equipment of the same type, said method comprising: Step a, extracting the historical data of the N electric equipments in a period of time, obtaining K defect categories satisfying predetermined conditions according to the statistics of the historical data, and using the obtained K defect categories as defect analysis variables; Wherein, K and N are both positive integers; Step b. In the historical data, for each defect analysis variable, accumulate the duration of the N electric equipment under the same defect analysis variable, and accumulate K pieces of the N electric equipment under the same defect analysis variable The total duration is obtained by the following duration, and the initial probability of each defect state variable is determined according to the accumulated total duration and the accumulated duration corresponding to each defect analysis variable, and further determined by the K initial probabilities are combined into a vector P(0); Step c. Obtain the total number of occurrences of the K defect analysis variables respectively in the historical data, divide the time period into M time slices, and sequentially determine the N power values between adjacent time slices The equipment corresponds to the number of conversion modes of K*K defect analysis variables, and according to the total number of appearances of K defect analysis variables acquired and the number of conversion modes of K*K defect analysis variables, determine the K*K types The conversion probabilities corresponding to the conversion modes of the defect analysis variables are further combined into a matrix P from the K*K conversion probabilities; wherein, M is a positive integer; Step d. Determine the maximum value in P(1) according to the formula P(1)=P(0)*P, and use the defect category corresponding to the determined maximum value as the predicted defect before the new electric equipment is put into use. 2. The method of claim 1, further comprising: Obtain the defects currently occurring in a power equipment that has been put into use, and find the maximum value corresponding to the K conversion methods in the matrix P according to the type of the current defect, and assign the maximum value corresponding to the found value to The defect category is used as the next predicted defect of the current defect-occurring electrical equipment. 3. the method for claim 1 is characterized in that, the specific steps of described step a comprise: Extracting the historical data of the N electrical equipment in a time period, and determining the total number of occurrences of each defect in the historical data; Arrange the total number of occurrences of each defect in order from large to small, screen out the first K-1 defects with the largest total number of times, and use the category corresponding to the K-1 defects and the category without defects as the ones that meet the conditions K defect categories are further set as defect analysis variables. 4. the method for claim 1 is characterized in that, the specific steps of described step b comprise: In the historical data, the continuous occurrence time corresponding to each defect analysis variable of the N electric equipment is obtained, and the duration of the N electric equipment under the same defect analysis variable is screened out and accumulated to obtain K defect analysis The durations corresponding to the variables respectively; Accumulate the duration corresponding to the obtained K defect analysis variables respectively to obtain the total duration; Dividing the durations corresponding to the obtained K defect analysis variables respectively by the obtained total duration to obtain initial probabilities corresponding to the K defect analysis variables respectively, and combining the K initial probabilities into a vector P(0). 5. the method for claim 1 is characterized in that, the specific steps of described step c comprise: Obtaining the total number of occurrences of the K defect analysis variables respectively in the historical data; Perform pairwise mapping on the K defect analysis variables to obtain K*K defect analysis variable conversion modes; Dividing the time period into M time slices, sorting the time from small to large, and determining the number of conversion times of the N electrical equipment corresponding to K*K defect analysis variable conversion modes between adjacent time slices; The number of conversions of each defect analysis variable is used as the molecule, and the defect analysis variable corresponding to the main map in each molecule is determined, and the total number of occurrences of the defect analysis variable corresponding to the main map in each molecule is selected as Corresponding to the denominator, the conversion probabilities corresponding to the K*K defect analysis variable conversion modes are obtained, and the K*K conversion probabilities are further combined into a matrix P. 6. The method according to any one of claims 1 to 5, wherein the electrical equipment is a transformer, and the defect analysis variables include no defect, oil leakage, cooling system failure, instrument failure, operation Abnormal mechanism and external mechanical damage. 7. A method for predicting electrical equipment defects, characterized in that it is implemented on N electrical equipment of the same type, said method comprising: Step S1, extracting the historical data of the N electrical equipment in a period of time, and obtaining K defect categories satisfying predetermined conditions according to the statistics of the historical data, and using the obtained K defect categories as defect analysis variables; Wherein, K and N are both positive integers; Step S2. Obtain the total number of occurrences of the K defect analysis variables respectively in the historical data, divide the time period into M time slices, and sequentially determine the N power values between adjacent time slices The equipment corresponds to the number of conversion modes of K*K defect analysis variables, and according to the total number of appearances of K defect analysis variables acquired and the number of conversion modes of K*K defect analysis variables, determine the K*K types The conversion probabilities corresponding to the conversion modes of the defect analysis variables are further combined into a matrix P from the K*K conversion probabilities; wherein, M is a positive integer; Step S3. Obtain the current defect of a power equipment that has been put into use, and find the maximum value corresponding to the K conversion methods in the matrix P according to the type of the current defect, and convert the found maximum value to The defect category corresponding to the value is used as the next predicted defect of the electric equipment that currently has a defect. 8. The method according to claim 7, wherein the specific steps of the step S1 include: Extracting the historical data of the N electrical equipment in a time period, and determining the total number of occurrences of each defect in the historical data; Arrange the total number of occurrences of each defect in order from large to small, screen out the first K-1 defects with the largest total number of times, and use the category corresponding to the K-1 defects and the category without defects as the ones that meet the conditions K defect categories are further set as defect analysis variables. 9. The method according to claim 7, wherein the specific steps of the step S2 include: Obtaining the total number of occurrences of the K defect analysis variables respectively in the historical data; Perform pairwise mapping on the K defect analysis variables to obtain K*K defect analysis variable conversion modes; Dividing the time period into M time slices, sorting the time from small to large, and determining the number of conversion times of the N electrical equipment corresponding to K*K defect analysis variable conversion modes between adjacent time slices; The number of conversions of each defect analysis variable is used as the molecule, and the defect analysis variable corresponding to the main map in each molecule is determined, and the total number of occurrences of the defect analysis variable corresponding to the main map in each molecule is selected as Corresponding to the denominator, the conversion probabilities corresponding to the K*K defect analysis variable conversion modes are obtained, and the K*K conversion probabilities are further combined into a matrix P. 10. The method of claim 7, further comprising: In the historical data, for each defect analysis variable, the duration of the N electric equipment under the same defect analysis variable is accumulated, and K durations of the N electric equipment under the same defect analysis variable are accumulated time to obtain the total duration, and according to the accumulated total duration and the accumulated duration corresponding to each defect analysis variable, determine the initial probability of each defect state variable, and further from the K initial The probabilities are combined into a vector P(0); According to the formula P(1)=P(0)*P, the maximum value in P(1) is determined, and the defect category corresponding to the determined maximum value is used as the predicted defect before the new electric equipment is put into use. 11. The method according to claim 10, wherein, in the historical data, for each defect analysis variable, the duration of the N electric equipment under the same defect analysis variable is accumulated, and accumulating durations of K pieces of said N electrical equipment under the same defect analysis variable to obtain a total duration, and according to the accumulated total duration and the accumulated duration corresponding to each defect analysis variable, Determine the initial probability of each defect state variable, and the specific steps of further combining the K initial probabilities into a vector P (0) include: In the historical data, the continuous occurrence time corresponding to each defect analysis variable of the N electric equipment is obtained, and the duration of the N electric equipment under the same defect analysis variable is screened out and accumulated to obtain K defect analysis The durations corresponding to the variables respectively; Accumulate the duration corresponding to the obtained K defect analysis variables respectively to obtain the total duration; Dividing the durations corresponding to the obtained K defect analysis variables respectively by the obtained total duration to obtain initial probabilities corresponding to the K defect analysis variables respectively, and combining the K initial probabilities into a vector P(0). 12. A system for predicting defects in electrical equipment, characterized in that it is implemented on N electrical equipment of the same type, the system comprising: Determining a defect analysis variable unit for extracting the historical data of the N electrical equipment in a period of time, obtaining K defect categories satisfying predetermined conditions according to the statistics of the historical data, and dividing the obtained K defect categories As a defect analysis variable; among them, K and N are both positive integers; Acquiring an initial probability unit for accumulating the duration of the N electric equipment under the same defect analysis variable for each defect analysis variable in the historical data, and accumulating K pieces of the N electric equipment at The total duration is obtained from the duration under the same defect analysis variable, and the initial probability of each defect state variable is determined according to the accumulated total duration and the accumulated duration corresponding to each defect analysis variable, Further be combined into vector P (0) by described K initial probabilities; Obtaining a conversion probability unit for obtaining the total number of occurrences of the K defect analysis variables in the historical data, and dividing the time period into M time slices, and sequentially determining the number of times between adjacent time slices The N electric equipments correspond to the number of times of conversion of K*K kinds of defect analysis variables, and according to the total number of occurrences of the acquired K defect analysis variables and the number of K*K kinds of defect analysis variable conversion times, determine the The conversion probabilities corresponding to K*K kinds of defect analysis variable conversion modes are further combined into a matrix P by the K*K conversion probabilities; wherein, M is a positive integer; The new equipment prediction defect unit is used to determine the maximum value in P(1) according to the formula P(1)=P(0)*P, and put the defect category corresponding to the determined maximum value into use as new electric equipment previous prediction deficiencies. 13. The system of claim 12, further comprising: The operating equipment prediction defect unit is used to obtain the current defects of a power device that has been put into use, and find the maximum value corresponding to the K conversion methods in the matrix P according to the type of the current defect, and convert the The defect category corresponding to the found maximum value is used as the next predicted defect of the electric equipment that currently has a defect. 14. A system for predicting defects in electrical equipment, characterized in that it is implemented on N electrical equipment of the same type, said system comprising: Determining a defect analysis variable unit for extracting the historical data of the N electrical equipment in a period of time, obtaining K defect categories satisfying predetermined conditions according to the statistics of the historical data, and dividing the obtained K defect categories As a defect analysis variable; among them, K and N are both positive integers; Obtaining a conversion probability unit for obtaining the total number of occurrences of the K defect analysis variables in the historical data, and dividing the time period into M time slices, and sequentially determining the number of times between adjacent time slices The N electric equipments correspond to the number of times of conversion of K*K kinds of defect analysis variables, and according to the total number of occurrences of the acquired K defect analysis variables and the number of K*K kinds of defect analysis variable conversion times, determine the The conversion probabilities corresponding to K*K kinds of defect analysis variable conversion modes are further combined into a matrix P by the K*K conversion probabilities; wherein, M is a positive integer; The operating equipment prediction defect unit is used to obtain the current defects of a power device that has been put into use, and find the maximum value corresponding to the K conversion methods in the matrix P according to the type of the current defect, and convert the The defect category corresponding to the found maximum value is used as the next predicted defect of the electric equipment that currently has a defect. 15. The system of claim 14, further comprising: Acquiring an initial probability unit for accumulating the duration of the N electric equipment under the same defect analysis variable for each defect analysis variable in the historical data, and accumulating K pieces of the N electric equipment at The total duration is obtained from the duration under the same defect analysis variable, and the initial probability of each defect state variable is determined according to the accumulated total duration and the accumulated duration corresponding to each defect analysis variable, Further be combined into vector P (0) by described K initial probabilities; The new equipment prediction defect unit is used to determine the maximum value in P(1) according to the formula P(1)=P(0)*P, and put the defect category corresponding to the determined maximum value into use as new electric equipment previous prediction deficiencies.