Cable fault risk assessment method and system, terminal and medium

Abstract

The invention discloses a cable fault risk assessment method and system, a terminal and a medium. The method comprises the following steps: acquiring an operation record data set including cable influence factors and faults; dividing into common factors and rare factors according to the support degree; high-risk factors in the common factors and high-risk low-probability factors in the rare factors are screened out according to the confidence coefficient and the improvement degree; constructing a Bayesian model based on the influence factors and the cable fault result; determining the weight of each high-risk factor according to the coupling relationship between the high-risk factors; determining the weight of each high-risk low-probability factor according to the coupling relationship between each high-risk low-probability factor and each high-risk factor; and obtaining a cable fault risk assessment result according to the weights of the high-risk factors and the high-risk low-probability factors and the scores corresponding to the factors. On the basis of the Bayesian model, strong correlation factors and high-risk low-probability factors are considered, an influence factor weight calculation method is optimized, and a more accurate risk assessment result can be obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of cable risk assessment, and in particular to a cable fault risk assessment method, system, terminal and medium taking into account data coupling relationship and rare factors. Background Art

[0002] At present, the cable fault risk assessment schemes mainly include the following:

[0003] Huang Xinbo et al. proposed a study on the risk assessment model of cable lines based on fuzzy mathematics [J]. High Voltage Electrical Appliances, 2021, 57(09): 19-25. This method uses fuzzy mathematics to evaluate the weights of cable operating parameters and environmental factors, and then evaluates the cable risk coefficient. However, the weight allocation and factor comparison in the fuzzy analysis method often rely on the subjective judgment of the decision maker. The evaluation results are highly subjective, and the consistency of the evaluation matrix is ​​difficult to verify, making it difficult to ensure the accuracy and rationality of the decision-making basis.

[0004] Tan Xiao et al. proposed the application and comparison of Logistic regression and support vector machine in cable aging assessment [J]. High Voltage Electrical Appliances, 2023, 59(05): 113-121. This method uses the Logistic regression method to judge the impact of different influencing factors on cable failure. However, Logistic regression is prone to underfitting when dealing with complex data or when there is a strong correlation between features, resulting in reduced classification accuracy. For cable risk assessment involving multiple risk factors, its accuracy is not as good as the Bayesian method.

[0005] Liu Haikang et al. proposed a high-voltage cable reliability analysis method combining Bayesian inference with Weibull proportional risk model [J]. High Voltage Technology, 2021, 47(02): 546-554. This method uses a combination of traditional Bayesian model and Weibull proportional risk model to analyze the operational reliability of high-voltage cables. However, the traditional Bayesian method has poor classification effect when there is correlation between sample attributes, and the conditional probability of rare factors in sample data is very low, making it difficult to accurately assess the impact of rare factors on cable failure risk.

[0006] None of the existing methods consider the coupling relationship between the influencing factors and the impact of rare factors on the cable failure risk. There may be factors with strong coupling relationships among the influencing factors. If they are assumed to be independent in the evaluation model and the impact of each factor on the cable failure risk is calculated separately, the accuracy of the risk assessment will be reduced. There are some rare factors with very low frequency of occurrence among the influencing factors, and traditional methods often directly ignore the impact of these factors. However, there are also some factors among the rare factors that pose a great threat to cable safety. If their impact is ignored, the accuracy of the risk assessment results will be reduced. To this end, the present invention considers the coupling relationship of data and the impact of rare factors on the cable failure risk, and proposes a cable failure risk assessment scheme that takes into account the coupling relationship of data and rare factors. Summary of the invention

[0007] In view of the deficiencies in the prior art, the purpose of the present invention is to provide a cable fault risk assessment method, system, terminal and medium, which considers the coupling relationship of data and the influence of rare factors on cable fault risk, analyzes a large amount of cable operation and fault information to obtain prior information, calculates the posterior probability distribution through the Bayesian conditional probability formula, and then corrects the posterior probability in combination with the coupling relationship between factors, so as to obtain a more objective and accurate cable risk assessment result.

[0008] In a first aspect, a cable fault risk assessment method is provided, comprising the following steps:

[0009] Obtaining operation record data sets including cable influencing factors and faults;

[0010] Divide each influencing factor into common factors and rare factors according to the support;

[0011] Screen out high-risk factors among common factors and high-risk low-probability factors among rare factors based on confidence and lift;

[0012] Building a Bayesian model based on high-risk factors, high-risk low-probability factors, and cable failure results;

[0013] Determine the Bayesian weight coefficient of each high-risk factor based on the coupling relationship between high-risk factors; determine the Bayesian weight coefficient of each high-risk low-probability factor based on the coupling relationship between each high-risk low-probability factor and each high-risk factor;

[0014] According to the Bayesian weight coefficients of high-risk factors and high-risk low-probability factors and the corresponding scores of each factor, the cable fault risk assessment results are obtained.

[0015] Furthermore, each influencing factor is divided into common factors and rare factors according to the support, including:

[0016] The support of each influencing factor is calculated by the following formula:

[0017]

[0018] Where support(A) represents the support of influencing factor A; num[A∈X(i,j)] represents the number of running records containing influencing factor A in the running record dataset X(i,j), i and j represent the i-th row and j-th column respectively; num[X] represents the number of running records in the running record dataset;

[0019] A support threshold is set. When the support of an influencing factor is less than the support threshold, the influencing factor is considered to be a rare factor, otherwise it is a common factor.

[0020] Furthermore, high-risk factors among common factors are screened out based on confidence and lift, including:

[0021] The confidence and lift of each common factor are calculated using the following formula:

[0022]

[0023] Among them, confidence (A c →B) and lift(A c →B) respectively represent common factors A c Confidence and improvement of cable fault B; num[A c ∈X(i,j); b i =1] indicates that the running record data set X(i,j) contains common factors A c and is the number of running records of cable faults, b i =1 indicates that the operation record is a cable fault record, i and j represent the i-th row and j-th column respectively; num[A c ∈X(i,j)] means that the running record data set X(i,j) contains common factors A c The number of running records; support(B) represents the support of cable fault B, num[B∈X(i,j); b i =1] represents the number of operation records of cable faults in the operation record data set X(i,j], and num[X] represents the number of operation records in the operation record data set;

[0024] A first confidence threshold and a first lift threshold are set. When the confidence of a common factor is not less than the first confidence threshold and the lift of the common factor is not less than the first lift threshold, the common factor is considered to be a high-risk factor.

[0025] Furthermore, high-risk and low-probability factors among rare factors are screened out based on confidence and lift, including:

[0026] The confidence and lift of each rare factor are calculated using the following formula:

[0027]

[0028] Among them, confidence rare (A r →B) and lift rare (A r →B) respectively represent rare factors A r Confidence and improvement of cable fault B; num[A r ∈X(i,j);A c ∈X(i,j'); b i =1] indicates that the running record data set X(i,j) also contains rare factors A r and common factor A c and is the number of running records of cable faults, b i =1 indicates that the operation record is a cable fault record, i and j represent the i-th row and j-th column respectively; Represents the running record data set X containing rare factors r Contains rare factor A r The number of running records; Represents the running record data set X containing rare factors r Contains rare factor A r And the number of operation records of cable failures;

[0029] A second confidence threshold and a second lift threshold are set. When a rare factor exists and its corresponding confidence is not less than the second confidence threshold and its lift is not less than the second lift threshold, the rare factor is considered to be a high-risk low-probability factor.

[0030] Furthermore, the Bayesian weight coefficient of each high-risk factor is determined according to the coupling relationship between the high-risk factors, specifically including:

[0031] When a cable's operating record data set contains high risk factors If there are no other influencing factors with strong coupling relationship with it, the risk weight of this high risk factor is for:

[0032]

[0033] In the formula, is the high risk factor calculated according to the Bayesian conditional probability formula The posterior probability distribution of cable failure under certain circumstances;

[0034] When a cable's operating record data set contains high risk factors And there are other high-risk factors with strong coupling relationship with it The risk weight of the high risk factor is for:

[0035]

[0036] In the formula, is the high risk factor calculated according to the Bayesian conditional probability formula and At the same time, the posterior probability distribution of cable failure is determined, M is the factor with high risk The total number of other high-risk factors with strong coupling relationships.

[0037] Furthermore, the Bayesian weight coefficient of each high-risk low-probability factor is determined according to the coupling relationship between each high-risk low-probability factor and each high-risk factor, specifically including:

[0038] According to the confidence and lift, high-risk factors with strong coupling relationships with various high-risk low-probability factors are screened out;

[0039] For high risk and low probability factors Its risk weight for:

[0040]

[0041] In the formula, is the high risk factor calculated according to the Bayesian conditional probability formula High risk low probability factors occur in certain situations The posterior probability distribution of ; is the high risk factor calculated according to the Bayesian conditional probability formula The posterior probability distribution of cable failure under certain circumstances; Both are related to high risk and low probability factors There are high-risk factors with strong coupling relationships.

[0042] Furthermore, high-risk factors that have a strong coupling relationship with each high-risk low-probability factor are screened out based on confidence and lift, including:

[0043] The confidence of each high-risk low-probability factor about each high-risk factor is calculated by the following formula: rare (A r' →A c' ) and lift rare (A r' →A c' ):

[0044]

[0045] In the formula, num r [A c' ∈X(i,j);A r' ∈X(i,j')] represents the running record data set X containing all rare factors r Contains both high risk and low probability factors A r' and high risk factor A c' The number of running records; Represents all running record data sets X containing rare factors r Includes high risk and low probability factors A r' The number of running records; num r [A c' ∈X c ] represents the number of running records containing high-risk factors in the running record data set containing all rare factors, X c represents a running record data set containing common factors;

[0046] A third confidence threshold and a third lift threshold are set. When the confidence and lift of a high-risk low-probability factor with respect to a certain high-risk factor simultaneously satisfy no less than the corresponding third confidence threshold and third lift threshold, the high-risk factor is considered to have a strong coupling relationship with the high-risk low-probability factor.

[0047] In a second aspect, a cable fault risk assessment system is provided, comprising:

[0048] Data acquisition module, used to obtain operation record data sets including cable influencing factors and faults;

[0049] The factor classification module is used to classify each influencing factor into common factors and rare factors according to the support degree;

[0050] High-risk factor screening module, used to screen out high-risk factors among common factors and high-risk low-probability factors among rare factors based on confidence and lift;

[0051] A Bayesian model building module for building a Bayesian model based on high-risk factors, high-risk low-probability factors, and cable fault results;

[0052] The Bayesian weight coefficient calculation module is used to determine the Bayesian weight coefficient of each high-risk factor according to the coupling relationship between the high-risk factors; determine the Bayesian weight coefficient of each high-risk low-probability factor according to the coupling relationship between each high-risk low-probability factor and each high-risk factor;

[0053] The cable fault risk assessment module is used to obtain the cable fault risk assessment result according to the Bayesian weight coefficients of high-risk factors and high-risk low-probability factors and the scores corresponding to each factor.

[0054] In a third aspect, an electronic terminal is provided, comprising:

[0055] a memory having a computer program stored thereon;

[0056] The processor is used to load and execute the computer program to implement the cable fault risk assessment method as described above.

[0057] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the cable fault risk assessment method as described above is implemented.

[0058] The present invention proposes a cable fault risk assessment method, system, terminal and medium, which have the following beneficial effects:

[0059] (1) The present invention uses three importance diagnostic criteria, namely support, confidence and lift, to distinguish the influencing factors in the cable operation record data. For rare factors, an improved calculation method of the importance diagnostic criteria is proposed, and the high-risk and low-probability factors with a strong coupling relationship with cable faults are analyzed. Based on the Bayesian model, the strong correlation factors and high-risk and low-probability factors in the cable operation record data are mainly mined, and the posterior probability distribution calculation method of the model is optimized, thereby forming an optimized cable fault risk assessment method.

[0060] (2) Compared with the traditional cable risk assessment method, the present invention makes full use of the big data information of cable operation, combines the prior information of the cable, considers the coupling relationship between different influencing factors and the impact of rare factors on the cable failure risk, optimizes the calculation of weights in the Bayes model, and can obtain more accurate risk assessment results, providing a basis and guidance for fault warning and differentiated maintenance, so that the cable can evolve from post-fault maintenance to differentiated status maintenance, effectively reducing the cost of fault detection and prevention. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0062] Figure 1It is a flow chart of a cable fault risk assessment method provided by an embodiment of the present invention;

[0063] Figure 2 Schematic diagram of a Bayesian network model for cable fault risk assessment provided by an embodiment of the present invention;

[0064] Figure 3 is a comparison diagram of ROC curves of risk assessment results provided by an embodiment of the present invention;

[0065] Figure 4 It is a PR curve comparison of risk assessment results provided by an embodiment of the present invention;

[0066] Figure 5 It is a DET curve comparison of the risk assessment results provided by the embodiment of the present invention;

[0067] Figure 6 It is a comparison of the YI change curves of the risk assessment results provided by the embodiments of the present invention. DETAILED DESCRIPTION

[0068] To make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be described in detail below. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other implementation methods obtained by ordinary technicians in this field without creative work belong to the scope of protection of the present invention.

[0069] In the description of the present invention, the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or order. In addition, in the description of the present invention, unless otherwise specified, the meaning of "plurality" refers to at least two.

[0070] The present invention proposes a cable fault risk assessment scheme which takes into account data coupling relationship and rare factors. The scheme considers the influence of data coupling relationship and rare factors on cable fault risk, analyzes a large amount of cable operation and fault information to obtain prior information, calculates the posterior probability distribution through the Bayesian conditional probability formula, and then corrects the posterior probability in combination with the coupling relationship between factors, thereby obtaining a more objective and accurate cable risk assessment result.

[0071] Coupling relationship analysis is often used to reflect the potential impact relationship between different things, and can intuitively and flexibly analyze the deep connection between fault data and cable information. There are two main forms of coupling relationship: high-frequency variable set and frequent coupling relationship. High-frequency variable set refers to the variable set whose frequency exceeds the threshold, and frequent coupling relationship refers to the coupling relationship where the probability of two variables appearing together is greater than the threshold.

[0072] The traditional coupling relationship analysis algorithm (CRA) uses the same and fixed importance diagnosis standard and threshold setting method, which will cause the rare influencing factors in the input operation record data to be directly discarded without analysis. However, the method proposed in this paper needs to screen rare factors from the input data and analyze the coupling relationship between them and cable faults to mine the high-risk and low-probability factors. Therefore, it is necessary to improve the traditional CRA importance calculation formula.

[0073] A coupling relationship can be written as A→B, where A is a set of influencing factors and B is a target variable or a set of target variables. In order to analyze common factors and rare factors separately, this paper redefines a coupling relationship as A c +A r →B, where A c and A r They represent the set of common factors and the set of rare factors in the input operation record data respectively. This method uses three important diagnostic criteria of support, confidence and lift to perform data mining. The technical solution of the present invention is specifically described below in conjunction with specific embodiments.

[0074] like Figure 1 As shown, an embodiment of the present invention provides a cable fault risk assessment method, comprising the following steps:

[0075] S1: Obtain an operation record data set including cable influencing factors and faults.

[0076] The operation record data set can be listed in a list or matrix format, with each row being an operation record and each column of each operation record being different influencing factors and cable fault status data.

[0077] S2: Divide each influencing factor into common factors and rare factors according to the support.

[0078] In this embodiment, the support of each influencing factor is calculated by the following formula:

[0079]

[0080] Where support(A) represents the support of influencing factor A; num[A∈X(i,j)] represents the number of running records containing influencing factor A in the running record dataset X(i,j), i and j represent the i-th row and j-th column respectively; num[X] represents the number of running records in the running record dataset;

[0081] A support threshold is set. When the support of an influencing factor is less than the support threshold, the influencing factor is considered to be a rare factor, otherwise it is a common factor.

[0082] S3: Filter out high-risk factors among common factors and high-risk low-probability factors among rare factors based on confidence and lift.

[0083] Among the rare factors, there are some high-risk, low-probability factors that have a great impact on cable failures. At present, the traditional CRA algorithm uses the same and fixed importance calculation method and threshold as the common elements in the corresponding features for calculation, which will cause these high-risk, low-probability elements to be directly excluded, and it is impossible to accurately evaluate the impact of these factors on the risk of cable failure. Therefore, this embodiment uses different methods to calculate the importance diagnosis standard for common factors and rare factors, and especially proposes an improved calculation method for the importance diagnosis standard for rare factors. The specific process is as follows.

[0084] S31: Filter out high-risk factors among common factors based on confidence and lift, including:

[0085] S311: Calculate the confidence and lift of each common factor using the following formula:

[0086]

[0087] Among them, confidence (A c →B) and lift(A c →B) respectively represent common factors A c Confidence and improvement of cable fault B; num[A c ∈X(i,j); b i =1] indicates that the running record data set X(i,j) contains common factors A c and is the number of running records of cable faults, b i =1 indicates that the operation record is a cable fault record, i and j represent the i-th row and j-th column respectively; num[A c ∈X(i,j)] means that the running record data set X(i,j) contains common factors A c The number of running records; support(B) represents the support of cable fault B, num[B∈X(i,j); b i =1] represents the number of operation records of cable faults in the operation record data set X(i,j), and num[X} represents the number of operation records in the operation record data set;

[0088] S312: Setting a first confidence threshold and a first lift threshold. When the confidence of a common factor is not less than the first confidence threshold and the lift of the common factor is not less than the first lift threshold, the common factor is considered to be a high-risk factor.

[0089] S32: Filter out high-risk and low-probability factors among rare factors based on confidence and lift, including:

[0090] S321: Calculate the confidence and lift of each rare factor using the following formula:

[0091]

[0092] Among them, confidence rare (A r →B) and lift rare (A r →B) respectively represent rare factors A r Confidence and improvement of cable fault B; num[A r ∈X(i,j);A c ∈X(i,j'); b i =1] indicates that the running record data set X(i,j) also contains rare factors A r and common factor A c and is the number of running records of cable faults, b i =1 indicates that the operation record is a cable fault record, i and j represent the i-th row and j-th column respectively; Represents the running record data set X containing rare factors r Contains rare factor A r The number of running records; Represents the running record data set X containing rare factors r Contains rare factor A r And the number of operation records of cable failures;

[0093] S322: Setting a second confidence threshold and a second lift threshold. When a rare factor exists and its corresponding confidence is not less than the second confidence threshold and its lift is not less than the second lift threshold, the rare factor is considered to be a high-risk low-probability factor.

[0094] It should be noted that when calculating the confidence and lift of a rare factor, the rare factor A r With each common factor A c The confidence and improvement in the coupled case are as long as there is a common factor A c With rare factor A r If the confidence and improvement in the coupled case are not less than their respective thresholds, then the rare factor A is considered r It is a high risk and low probability factor.

[0095] S4: Construct a Bayesian model based on high-risk factors, high-risk low-probability factors, and cable failure results.

[0096] The use of Bayesian network models for cable fault risk assessment is an analytical method based on probabilistic inference. By constructing a Bayesian network model, it can effectively represent the conditional dependencies between variables, and can combine known prior knowledge with historical data to infer the posterior probability of unknown variables. The use of this method can help determine the risk of future cable failures, thereby helping to develop more effective maintenance strategies and preventive measures. This assessment method is particularly suitable for complex systems that need to deal with uncertainty and probabilistic reasoning, and can also update model parameters based on subsequent collected data. A Bayesian network structure model can be established based on the operation and fault records of the cable, such as Figure 2 The network model can clearly show the various factors that affect the risk of cable failure and their interrelationships, helping to better understand and predict the risk of cable failure.

[0097] Figure 2 In the above example, X represents the running record data set; A 1 ,A 2 ,…,A x Represent different types of influencing factor sets; A x1 ,A x2 ...A xz represents all the influencing factors contained in the influencing factor set of the xth type, B 1 and B 2 Represents cable failure and no failure respectively, A xz and B 1,2 The solid line between represents the prior distribution relationship between each influencing factor and the cable failure, and the dotted line represents the prior distribution relationship between each influencing factor and the cable failure. The posterior distribution probability can be calculated using the Bayesian formula:

[0098]

[0099] In the formula, ω(A αβ |B) indicates influencing factor A αβ The posterior probability distribution of cable failure under certain conditions is calculated based on the prior distribution of influencing factors and the sample likelihood function. αβ ) represents the influencing factor A obtained by analyzing the operation record data set αβ The prior distribution of P(B|A αβ ) represents the influencing factor A calculated by the likelihood function of this sample αβ By calculating formula (7), the influence weights of different influencing factors on the risk of cable failure calculated according to the traditional Bayes network model can be obtained.

[0100] Bayesian networks have significant advantages when considering the coupling relationship between influencing factors. It mainly uses local distribution (conditional probability distribution) to describe joint distribution, and obtains global variable relationships by integrating the local interaction relationship between variables. This enables Bayesian networks to handle fuzziness and uncertainty problems, especially when the relationship between influencing factors is complex and difficult to quantify using traditional methods. The operating environment of power cables is complex, and their aging failures are more affected by various intricate environmental factors, which is particularly suitable for risk assessment using Bayesian networks.

[0101] S5: Determine the Bayesian weight coefficient of each high-risk factor based on the coupling relationship between high-risk factors; determine the Bayesian weight coefficient of each high-risk low-probability factor based on the coupling relationship between each high-risk low-probability factor and each high-risk factor. Specifically, it includes the following process:

[0102] S51: Determine the Bayesian weight coefficient of each high-risk factor according to the coupling relationship between the high-risk factors, specifically including:

[0103] S511: When a cable operation record data set contains high risk factors If there are no other influencing factors with strong coupling relationship with it, the risk weight of this high risk factor is for:

[0104]

[0105] In the formula, is the high risk factor calculated according to the Bayesian conditional probability formula The posterior probability distribution of cable failure under certain conditions.

[0106] S512: When a cable operation record data set contains high risk factors And there are other high-risk factors with strong coupling relationship with it The risk weight of the high risk factor is for:

[0107]

[0108] In the formula, is the high risk factor calculated according to the Bayesian conditional probability formula and At the same time, the posterior probability distribution of cable failure is determined, M is the factor with high risk The total number of other high-risk factors with strong coupling relationships.

[0109] In the traditional Bayesian network model, due to the large proportion of common factors in the data set, the calculation of the posterior distribution probability is mainly affected by the common factors. Rare factors appear very rarely in the operation record data and are often ignored without analysis, making it difficult to accurately calculate their posterior probability distribution with the target quantity. Therefore, this paper improves the calculation method of the posterior probability distribution of high-risk low-probability factors, thereby improving the accuracy and reliability of risk assessment.

[0110] S52: Determine the Bayesian weight coefficient of each high-risk low-probability factor according to the coupling relationship between each high-risk low-probability factor and each high-risk factor, specifically including:

[0111] S521: Filter out high-risk factors that have a strong coupling relationship with each high-risk low-probability factor based on confidence and lift.

[0112] Specifically, the confidence of each high-risk low-probability factor with respect to each high-risk factor is calculated by the following formula: rare (A r' →A c' ) and lift rare (A r' →A c' ):

[0113]

[0114] In the formula, num r [A c' ∈X(i,j);A r' ∈X(i,j')] represents the running record data set X containing all rare factors r Contains both high risk and low probability factors A r' and high risk factor A c' The number of running records; Represents all running record data sets X containing rare factors r Includes high risk and low probability factors A r' The number of running records; num r [A c' ∈X c ] represents the number of running records containing high-risk factors in the running record data set containing all rare factors, X c represents a running record data set containing common factors;

[0115] A third confidence threshold and a third lift threshold are set. When the confidence and lift of a high-risk low-probability factor with respect to a certain high-risk factor simultaneously satisfy no less than the corresponding third confidence threshold and third lift threshold, the high-risk factor is considered to have a strong coupling relationship with the high-risk low-probability factor.

[0116] S522: For high risk and low probability factors Its risk weight for:

[0117]

[0118] In the formula, is the high risk factor calculated according to the Bayesian conditional probability formula High risk low probability factors occur in certain situations The posterior probability distribution of ; is the high risk factor calculated according to the Bayesian conditional probability formula The posterior probability distribution of cable failure under certain circumstances; Both are related to high risk and low probability factors There are high-risk factors with strong coupling relationships.

[0119] S6: The cable fault risk assessment result is obtained based on the Bayesian weight coefficients of the high-risk factors and the high-risk low-probability factors and the scores corresponding to each factor.

[0120] It should be noted that how to optimize the weights of each influencing factor to make the final cable risk assessment result better is the focus of the present invention, and the scoring rules for each influencing factor are not the focus of the present invention. The existing scoring rule system can be used for this purpose, which will not be elaborated on here. In this embodiment, the final cable fault risk assessment result can be obtained by multiplying the scores of each high-risk factor and high-risk low-probability factor by their corresponding Bayesian weight coefficients and then summing them up.

[0121] The cable fault risk assessment method provided in the above embodiment uses three importance diagnostic criteria of support, confidence and lift to distinguish the influencing factors in the cable operation record data, and proposes an improved calculation method of the importance diagnostic criteria for rare factors, and analyzes the high-risk and low-probability factors with a strong coupling relationship with the cable fault in the rare factors. Based on the Bayes model, the strong correlation factors and high-risk and low-probability factors in the cable operation record data are mainly mined, and the posterior probability distribution calculation method of the model is optimized, thereby forming an optimized cable fault risk assessment method. Compared with the traditional cable risk assessment method, the present invention makes full use of the big data information of cable operation, combines the prior information of the cable, considers the coupling relationship between different influencing factors and the influence of rare factors on the cable fault risk, optimizes the calculation of weights in the Bayes model, and can obtain more accurate risk assessment results, providing a basis and guidance for fault warning and differentiated maintenance, so that the cable evolves from post-fault maintenance to differentiated state maintenance, and effectively reduces the cost of fault detection and prevention.

[0122] An embodiment of the present invention provides a cable fault risk assessment system, including:

[0123] Data acquisition module, used to obtain operation record data sets including cable influencing factors and faults;

[0124] The factor classification module is used to classify each influencing factor into common factors and rare factors according to the support degree;

[0125] High-risk factor screening module, used to screen out high-risk factors among common factors and high-risk low-probability factors among rare factors based on confidence and lift;

[0126] A Bayesian model building module for building a Bayesian model based on high-risk factors, high-risk low-probability factors, and cable fault results;

[0127] The Bayesian weight coefficient calculation module is used to determine the Bayesian weight coefficient of each high-risk factor according to the coupling relationship between the high-risk factors; determine the Bayesian weight coefficient of each high-risk low-probability factor according to the coupling relationship between each high-risk low-probability factor and each high-risk factor;

[0128] The cable fault risk assessment module is used to obtain the cable fault risk assessment result according to the Bayesian weight coefficients of high-risk factors and high-risk low-probability factors and the assessment points corresponding to each factor.

[0129] It should be understood that the functional unit modules in various embodiments of the present invention may be concentrated in one processing unit, or each unit module may exist physically separately, or two or more unit modules may be integrated in one unit module, and may be implemented in the form of hardware or software.

[0130] An embodiment of the present invention further provides an electronic terminal, including:

[0131] a memory having a computer program stored thereon;

[0132] The processor is used to load and execute the computer program to implement the cable fault risk assessment method described in the above embodiment.

[0133] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the cable fault risk assessment method described in the above embodiment is implemented.

[0134] It can be understood that the same or similar parts of the above embodiments can be referenced to each other, and the contents not described in detail in some embodiments can refer to the same or similar contents in other embodiments.

[0135] The following model is verified and tested using the operation records of 1797 distribution cables in a southern region in 2023. The corresponding high-risk factors and high-risk low-probability factors are analyzed by the coupling relationship analysis method proposed in this invention, where the thresholds corresponding to the support, confidence and lift are set as minsupport, 0 =0.2,minconfidence 0 =0.5,minlift 0 =1.1. The influencing factors included in the test data are shown in Table 1.

[0136] Table 1 Summary of the influencing factors studied

[0137]

[0138] To verify the accuracy of the risk assessment results, the present invention uses the receiver operating characteristic curve (ROC), precision-recall curve (PR) and detection error trade-off ratio curve (DET) to evaluate the performance of the risk assessment model. At the same time, the area under the ROC curve (Area Under the ROC, AUROC), the area under the PR curve (Area Under the PR, AUPR), the equal error rate (EER) and Youden's Index (YI) are used as evaluation indicators to more intuitively reflect the effect of the risk assessment model. AUROC and AUPR reflect the prediction effect of the model, and the larger the values ​​of the two, the better the effect. EER reflects the exclusion rate and error rate of the model, and the smaller its value, the better the model effect. YI reflects the authenticity of the experimental event, and the larger its value, the better the model effect.

[0139] In the cable operation and fault records collected by the present invention, the cable fault results are divided into two types: faults occurred and no faults occurred. If the risk score of a cable operation record is high, the failure result of the cable is likely to be a failure. The present invention uses the proposed cable fault risk assessment method that takes into account data coupling relationships and rare factors to conduct risk assessment on cable operation records. In order to verify the advantages of the method proposed in the present invention, the evaluation results of the traditional Bayes method and Logistic regression method are added for comparison. The test ROC curve, PR curve, DET curve and Youden index change curve of the cable fault risk assessment results of the method proposed in the present invention (optimized Bayes) and the traditional Bayes model and Logistic regression model are as shown in Figure 3 , 4, 5, and 6. The results of relevant evaluation parameters are shown in Table 2.

[0140] Table 2 Comparison of evaluation results

[0141]

[0142] Depend on Figure 3 , Figure 4 , Figure 5 , Figure 6 As shown in Table 2, compared with the traditional Bayesian model and the Logistic regression model, the optimized Bayesian model proposed in the present invention can obtain better prediction results. Compared with the traditional Bayesian model and the Logistic regression model, the method proposed in the present invention improves the scores of AUROC, AUPR, YI and EER by 9.8% and 4.6%, 8.5% and 5.3%, 15.6% and 7.8%, 6.3% and 3.8%, respectively, which shows that the optimized Bayesian model has a better performance effect, and the risk assessment results have better credibility and authenticity.

[0143] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0144] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0145] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0146] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0147] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes additional implementations, in which the order shown or discussed may not be followed, including performing functions in a substantially simultaneous manner depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention belong.

[0148] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A cable fault risk assessment method, characterized in that: The steps include: Obtaining operation record data sets including cable influencing factors and faults; Divide each influencing factor into common factors and rare factors according to the support; Filter out high-risk factors among common factors and high-risk low-probability factors among rare factors based on confidence and lift; Building a Bayesian model based on high-risk factors, high-risk low-probability factors, and cable failure results; Determine the Bayesian weight coefficient of each high-risk factor according to the coupling relationship between the high-risk factors; Determine the Bayesian weight coefficient of each high-risk low-probability factor according to the coupling relationship between each high-risk low-probability factor and each high-risk factor; According to the Bayesian weight coefficients of high-risk factors and high-risk low-probability factors and the corresponding scores of each factor, the cable fault risk assessment results are obtained.

2. The cable fault risk assessment method according to claim 1, characterized in that: According to the support, each influencing factor is divided into common factors and rare factors, including: The support of each influencing factor is calculated by the following formula: Where support(A) represents the support of influencing factor A; num[A∈X(i, j)] represents the number of running records containing influencing factor A in the running record dataset X(i, j), i and j represent the i-th row and j-th column respectively; num[X] represents the number of running records in the running record dataset; A support threshold is set. When the support of an influencing factor is less than the support threshold, the influencing factor is considered to be a rare factor, otherwise it is a common factor.

3. The cable fault risk assessment method according to claim 1, characterized in that: High-risk factors among common factors are screened out based on confidence and lift, including: The confidence and lift of each common factor are calculated using the following formula: Among them, confidence (A c →B) and lift(A c →B) respectively represent common factors A c Confidence and improvement of cable fault B; num[A c ∈X(i,j); b i =1] indicates that the running record data set X(i, j) contains common factors A c and is the number of running records of cable faults, b i =1 indicates that the operation record is a cable fault record, i and j represent the i-th row and j-th column respectively; num[A c ∈X(i, j)] means that the running record data set X(i, j) contains common factors A c The number of running records; support(B) represents the support of cable fault B, num[B∈X(i, j); b i =1] represents the number of operation records of cable faults in the operation record data set X(i, j), and num[X] represents the number of operation records in the operation record data set; A first confidence threshold and a first lift threshold are set. When the confidence of a common factor is not less than the first confidence threshold and the lift of the common factor is not less than the first lift threshold, the common factor is considered to be a high-risk factor.

4. The cable fault risk assessment method according to claim 1, characterized in that: High-risk and low-probability factors among rare factors are screened out based on confidence and lift, including: The confidence and lift of each rare factor are calculated using the following formula: Among them, confidence rare (A r →B) and lfft rare (A r →B) respectively represent rare factors A r Confidence and improvement of cable fault B; num[A r ∈X(i, j); A c ∈X(i, j′); b i =1] indicates that the running record data set X(i, j) also contains rare factors A r and common factor A c and is the number of running records of cable faults, b i =1 indicates that the operation record is a cable fault record, i and j represent the i-th row and j-th column respectively; Represents the running record data set X containing rare factors r Contains rare factor A r The number of running records; Represents the running record data set X containing rare factors r Contains rare factor A r And the number of operation records of cable failures; A second confidence threshold and a second lift threshold are set. When a rare factor exists and its corresponding confidence is not less than the second confidence threshold and its lift is not less than the second lift threshold, the rare factor is considered to be a high-risk low-probability factor.

5. The cable fault risk assessment method according to claim 1, characterized in that: The Bayesian weight coefficient of each high-risk factor is determined according to the coupling relationship between the high-risk factors, including: When a cable's operating record data set contains high risk factors If there are no other influencing factors with strong coupling relationship with it, the risk weight of this high risk factor is for: In the formula, is the high risk factor calculated according to the Bayesian conditional probability formula The posterior probability distribution of cable failure under certain circumstances; When a cable's operating record data set contains high risk factors And there are other high-risk factors with strong coupling relationship with it The risk weight of the high risk factor is for: In the formula, is the high risk factor calculated according to the Bayesian conditional probability formula and At the same time, the posterior probability distribution of cable failure is determined, M is the factor with high risk The total number of other high-risk factors with strong coupling relationships.

6. The cable fault risk assessment method according to claim 1, characterized in that: The Bayesian weight coefficient of each high-risk low-probability factor is determined according to the coupling relationship between each high-risk low-probability factor and each high-risk factor, specifically including: According to the confidence and lift, high-risk factors with strong coupling relationships with various high-risk low-probability factors are screened out; For high risk and low probability factors Its risk weight for: In the formula, is the high risk factor calculated according to the Bayesian conditional probability formula High risk low probability factors occur in certain situations The posterior probability distribution of ; is the high risk factor calculated according to the Bayesian conditional probability formula The posterior probability distribution of cable failure under certain circumstances; Both are related to high risk and low probability factors There are high-risk factors with strong coupling relationships.

7. The cable fault risk assessment method according to claim 6, characterized in that: According to the confidence and lift, high-risk factors with strong coupling relationships with various high-risk low-probability factors are screened out, including: The confidence of each high-risk low-probability factor about each high-risk factor is calculated by the following formula: rare (A r′ →A c′ ) and lift rare (A r′ →A c′ ): In the formula, num r [A c′ ∈X(i, j); A r′ ∈X(i, j′)] represents the running record data set X containing all rare factors r Contains both high risk and low probability factors A r′ and high risk factor A c′ The number of running records; Represents all running record data sets X containing rare factors r Includes high risk and low probability factors A r′ The number of running records; num r [A c′ ∈X c ] represents the number of running records containing high-risk factors in the running record data set containing all rare factors, X c represents a running record data set containing common factors; A third confidence threshold and a third lift threshold are set. When the confidence and lift of a high-risk low-probability factor with respect to a certain high-risk factor simultaneously satisfy no less than the corresponding third confidence threshold and third lift threshold, the high-risk factor is considered to have a strong coupling relationship with the high-risk low-probability factor.

8. A cable fault risk assessment system, characterized in that: include: Data acquisition module, used to obtain operation record data sets including cable influencing factors and faults; The factor classification module is used to classify each influencing factor into common factors and rare factors according to the support degree; High-risk factor screening module, used to screen out high-risk factors among common factors and high-risk low-probability factors among rare factors based on confidence and lift; A Bayesian model building module for building a Bayesian model based on high-risk factors, high-risk low-probability factors, and cable fault results; A Bayesian weight coefficient calculation module is used to determine the Bayesian weight coefficient of each high-risk factor according to the coupling relationship between the high-risk factors; Determine the Bayesian weight coefficient of each high-risk low-probability factor according to the coupling relationship between each high-risk low-probability factor and each high-risk factor; The cable fault risk assessment module is used to obtain the cable fault risk assessment result according to the Bayesian weight coefficients of high-risk factors and high-risk low-probability factors and the scores corresponding to each factor.

9. An electronic terminal, characterized in that: include: a memory having a computer program stored thereon; A processor is used to load and execute the computer program to implement the cable fault risk assessment method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the cable fault risk assessment method according to any one of claims 1 to 7 is implemented.

Images