Electric power operation risk management and control method, system and device and storage medium

Abstract

The invention discloses an electric power operation risk management and control method, system and device and a storage medium. The method comprises the following steps: analyzing electric power work human injury report records to form an electric power work human injury ontology knowledge base; according to the ontology knowledge base, a concept set, a relation set, an instance set and an axiom set are defined based on the ontology model, and an electric power operation human injury ontology model is obtained; obtaining a target case, calculating the similarity between the target case and each case in the case library based on a gray algorithm, and obtaining a case closest to the target case in the case library according to the similarity as a similar case; and generating an auxiliary management and control scheme for the accident associated with the target case according to the similar case obtained through retrieval and matching. According to the invention, ontology and case-based reasoning technologies are applied to electric power operation risk management and control, and powerful support is provided for risk management and control of electric power construction site accidents.

Description

Technical Field

[0001] The present invention relates to the field of power operation risk control, and particularly to a power operation risk control method, system, device and storage medium. Background Art

[0002] As an important part of the country's infrastructure, the safe operation of power construction sites plays a crucial role in ensuring the development of the national economy and people's livelihood. With the continuous growth of power demand, the scale of power infrastructure projects is expanding, and the safety management and risk control capabilities during the construction process are facing severe challenges. Accidents at power construction sites are sudden and complex, requiring on-site management personnel to quickly and accurately conduct accident risk control and make correct risk control decisions in a timely manner to ensure personnel safety and project quality.

[0003] Currently, the risk control response at power construction sites mainly relies on manual experience and emergency plans, which have certain limitations. The unpredictability of risks and the variability of on-site conditions pose many challenges to traditional risk control methods. In addition, the existing emergency system is not yet perfect in terms of emergency mechanisms and system construction, and it is difficult to meet the requirements of rapid and efficient risk control. Summary of the Invention

[0004] The purpose of this part is to outline some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Simplifications or omissions may be made in this part, as well as in the abstract and title of the present invention, to avoid obscuring the purpose of this part, the abstract and the title, and such simplifications or omissions shall not be used to limit the scope of the present invention.

[0005] In view of the problems existing in the above-mentioned prior art, the present invention is proposed. The present invention provides a power operation risk control method, system, device and storage medium to solve the problems in the field of power grid operation risk control, such as numerous and messy risk control records, diverse types of risk control measures, and difficult and rapid determination of risk control decisions.

[0006] To solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, a power operation risk control method is provided, and the method includes the following steps:

[0008] Analyze the records of power operation personal accident reports to form a power operation personal accident ontology knowledge base;

[0009] Based on the power operation personal accident ontology knowledge base, define a concept set, a relationship set, an instance set, and an axiom set based on the ontology model, and obtain a power operation personal accident ontology model according to the concept set, the relationship set, the instance set, and the axiom set;

[0010] Obtain the target case, calculate the similarity between the target case and each case in the case base based on the grey algorithm, and obtain the case in the case base that is closest to the target case as the similar case according to the similarity.

[0011] Generate an auxiliary control plan for the accident associated with the target case according to the retrieved and matched similar cases.

[0012] Furthermore, the concept set includes accidents, accident-causing factors, accident levels, and accident types, and each concept contains basic attributes and attribute types; the relationship set includes the relationships between part and whole, inheritance between concepts, instance and concept, and attribute and concept; the instance set is the instance set of concepts, which is used to describe the specific data of actual power accidents; the axiom set includes the value-taking rules of categorical variables and the normalization requirements for concept value-taking, and the rules are determined based on the experience of power safety accident handling and the characteristics of historical accident data.

[0013] Furthermore, calculating the similarity between the target case and each case in the case base based on the grey algorithm, and obtaining the case in the case base that is closest to the target case as the similar case includes:

[0014] Calculate the local similarity between the target case and the historical cases, including attribute similarity and structural similarity. Different calculation methods are used for numerical and symbolic attributes for the attribute similarity, and the structural similarity is calculated according to the relationship between the intersection and union of the concept attributes of the target case and each case in the case base.

[0015] Adopt the grey algorithm, and calculate the similarity between two cases in terms of causal concepts through the definitions of tree distance similarity, WUP similarity, and concept similarity.

[0016] According to the different weights of the attribute indicators, sum the weighted attribute similarities, and combine the structural similarity and the similarity in causal concepts to obtain the global similarity.

[0017] Obtain the case with the highest global similarity in the case base as the similar case of the target case.

[0018] Furthermore, different calculation methods are used for numerical and symbolic attributes for the attribute similarity, including:

[0019] For numerical attributes, calculate using the following formula:

[0020]

[0021] In the formula, case X * is the target case, X i represents the i-th case in the case base, max jRepresents the maximum value of the j-th characteristic attribute; min j Represents the minimum value of the j-th characteristic attribute; Represents the value of the j-th characteristic attribute of the target case, x ij Represents the value of the j-th characteristic attribute of the i-th case;

[0022] For symbolic attributes: When the characteristic attribute is an unordered enumerated variable, the following formula is used for calculation:

[0023]

[0024] When the characteristic attribute is an ordered enumerated variable, the following formula is used for calculation:

[0025]

[0026] Among them, count j Represents the number of values of the j-th characteristic attribute.

[0027] Furthermore, the structural similarity is calculated according to the following formula:

[0028]

[0029] Among them, W Q∩C Is the number of elements in the intersection of the concept attributes of the target case and the i-th case, W Q∪C Is the number of elements in the union of the concept attributes of the target case and the i-th case, w i Is the weight of the i-th attribute in the intersection, w k Is the weight of the k-th attribute in the union, m is the value of W Q∩C The value of, l is the value of W Q∪C The value of.

[0030] Furthermore, using the grey algorithm, through the definitions of tree distance similarity, WUP similarity, and concept similarity, calculate the similarity between two cases in terms of causal concepts, including:

[0031] Let the target case X * Contain m causal concepts {p 11 , p 12 ,..., p 1m}, the i-th case X in the case base i Contains n causal concepts, and denote the causal concept similarity between case X * And X i As Sim o (X * , X i ), the calculation steps are as follows:

[0032] I. With Sim t(c 1 ,c 2 ) represents the tree similarity of concepts c 1 and c 2 , denoted as Sim s (c 1 ,c 2 ); the WUP similarity between the upper-level concepts of c 1 and c 2 is denoted as Sim 1 and c 2 The concept similarity Sim c (c 1 ,c 2 ) is calculated by the formula ; according to the above calculation method, the similarity matrix between X * and X i is obtained: SimM ij = Sim c (c i ,c j );

[0033] II. Find the maximum similarity sequence, take the maximum value λ jz in the similarity matrix, and delete the row and column where the maximum value is located to obtain the matrix S′, and repeat this step until the matrix is empty; record the maximum value λ iz obtained each time as where t 0 = min(m,n); add the maximum value obtained each time to the maximum similarity sequence , and calculate the weight w i of the existing factors using the gray algorithm;

[0034] III.. Calculate the cause concept similarity according to the following formula: where B is a regulation parameter, and Sim stru (X * ,X i ) is the structural similarity between two cases.

[0035] Furthermore, the global similarity is calculated as follows:

[0036]

[0037] where w j represents the weight of the jth attribute, and Sim j (X * ,X i ) represents the attribute similarity of case X i and the target case X * in the jth attribute, and Sim stru (X* , X i ) represents case X i The structural similarity between case X and the target case X * is Sim o (X * , X i ) represents case X i The causal concept similarity between case X and the target case X * is as follows

[0038] Furthermore, the method further includes: after generating an auxiliary control plan each time, storing it in the case library as a new case to realize automatic update of the case library

[0039] In a second aspect, a power operation risk control system is provided, and the system includes

[0040] An ontology knowledge base construction module, configured to analyze power operation personal accident report records to form a power operation personal accident ontology knowledge base

[0041] An ontology model construction module, configured to define a concept set, a relationship set, an instance set, and an axiom set based on the power operation personal accident ontology knowledge base according to an ontology model, and obtain a power operation personal accident ontology model according to the concept set, the relationship set, the instance set, and the axiom set

[0042] A case reasoning module, configured to obtain a target case, calculate the similarity between the target case and each case in the case library based on a grey algorithm, and obtain the case in the case library that is closest to the target case as a similar case according to the similarity

[0043] A risk control assistance module, configured to generate an auxiliary control plan for an accident associated with the target case according to the retrieved and matched similar cases

[0044] Furthermore, the concept set includes accidents, accident-causing factors, accident levels, and accident types, and each concept includes basic attributes and attribute types; the relationship set includes the relationships between parts and wholes, inheritance between concepts, instances and concepts, and attributes and concepts; the instance set is an instance set of concepts, used to describe the specific data of actual power accidents; the axiom set includes the value-taking rules of categorical variables and the normalization requirements for concept value-taking, and the rules are determined based on power safety accident handling experience and historical accident data characteristics

[0045] Furthermore, the case reasoning module includes

[0046] The local similarity calculation unit is used to calculate the local similarity between the target case and the historical cases, including the attribute similarity and the structure similarity. Different calculation methods are adopted for the numerical and symbolic attributes for the attribute similarity, and the structure similarity is calculated according to the relationship between the intersection and union of the concept attributes of the target case and each case in the case library;

[0047] The cause concept similarity calculation unit is used to calculate the similarity between two cases in the cause concept by using the grey algorithm through the definitions of tree distance similarity, WUP similarity and concept similarity;

[0048] The global similarity calculation unit is used to weighted sum the attribute similarities according to the different weights of the attribute indicators, and combine the structure similarity and the similarity in the cause concept to obtain the global similarity;

[0049] The similar case determination unit is used to obtain the case with the highest global similarity in the case library as the similar case of the target case.

[0050] Further, different calculation methods are adopted for the numerical and symbolic attributes for the attribute similarity, including:

[0051] For numerical attributes, the following formula is used for calculation:

[0052]

[0053] In the formula, case X * is the target case, and X i represents the i-th case in the case library, and max j represents the maximum value of the j-th characteristic attribute; min j represents the minimum value of the j-th characteristic attribute; represents the value of the j-th characteristic attribute of the target case, and x ij represents the value of the j-th characteristic attribute of the i-th case;

[0054] For symbolic attributes: when the characteristic attribute is an unordered enumerated variable, the following formula is used for calculation:

[0055]

[0056] When the characteristic attribute is an ordered enumerated variable, the following formula is used for calculation:

[0057]

[0058] Among them, count j represents the number of values of the j-th characteristic attribute.

[0059] Further, the structure similarity is calculated according to the following formula:

[0060]

[0061] where W Q∩C is the number of elements in the intersection of the target case and the conceptual attributes of the i-th case, and W Q∪C is the number of elements in the union of the target case and the conceptual attributes of the i-th case, w i is the weight of the i-th attribute in the intersection, and w k is the weight of the k-th attribute in the union. m is the value of W Q∩C and l is the value of W Q∪C .

[0062] Furthermore, using the grey algorithm, the similarity between two cases in terms of causal concepts is calculated through the definitions of tree distance similarity, WUP similarity, and concept similarity, including:

[0063] Let the target case X * contain m causal concepts {p 11 , p 12 ,..., p 1m}, and the i-th case X i in the case base contain n causal concepts. Denote the causal concept similarity between cases X * and X i as Sim o (X * , X i ). The calculation steps are as follows:

[0064] I. Let Sim t (c 1 , c 2 ) represent the tree similarity between concepts c 1 and c 2 , and let Sim s (c 1 , c 2 ) represent the WUP similarity between the upper concepts of c 1 and c 2 . Then the concept similarity Sim 1 (c 2 , c c ) between concepts c 1 and c 2 is calculated by the formula ; According to the above calculation method, the similarity matrix between X * and X i is obtained: SimM ij = Sim c (c i , c j );

[0065] II. Find the sequence with the maximum similarity, and take the maximum value λ in the similarity matrix ij , and delete the row and column where the maximum value is located to obtain the matrix S′. Repeat this step until the matrix is empty; record the maximum value λ obtained each time ij as where t 0 = min(m, n); add the maximum value obtained each time to the sequence of maximum similarities , and calculate the weight w of the existing factors using the grey algorithm i ;

[0066] III. Calculate the similarity of cause concepts according to the following formula: where B is a regulation parameter, and Sim stru (X * , X i ) is the structural similarity between two cases

[0067] Furthermore, the global similarity is calculated as follows:

[0068]

[0069] where w j represents the weight of the j-th attribute, Sim j (X * , X i ) represents the attribute similarity of case X i and the target case X * in the j-th attribute, Sim stru (X * , X i ) represents the structural similarity between case X i and the target case X * , and Sim o (X * , X i ) represents the similarity of cause concepts between case X i and the target case X * . Thirdly, a computer device is provided. The device includes one or more processors; a memory; and one or more programs, where the one or more programs are stored in the memory and are configured to be executed by the one or more processors. When the program is executed by the processor, the steps of the power operation risk control method based on grey algorithm similarity matching as described in the first aspect of the present invention are implemented

[0070] In a fourth aspect, a computer storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the power operation risk control method based on grey algorithm similarity matching as described in the first aspect of the present invention are implemented.

[0071] Compared with the prior art, the beneficial effects of the invention are as follows: The present invention applies ontology and case-based reasoning technology to power operation risk control, providing strong support for the risk control of power construction site accidents. By introducing the grey algorithm into the calculation of global similarity, the effectiveness and convenience of case matching can be effectively improved, which is conducive to the standardized representation of risk control knowledge and knowledge sharing, as well as the improvement of power operation risk control efficiency and accuracy. On the other hand, the system constructed by the present invention can realize the rapid retrieval and matching of similar cases through case-based reasoning, effectively assisting the formation of risk control measure plans and improving the scientificity, systematicness and real-time nature of power operation risk control. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 is the overall flowchart of the method of the present invention;

[0073] Figure 2 is an example of accident case knowledge of the present invention;

[0074] Figure 3 is the flowchart of case similarity calculation of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0075] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0076] Embodiment 1

[0077] This embodiment provides a power operation risk control method. Referring to Figure 1 , the method mainly includes three stages: the construction of the ontology model for power operation risk control, the case-based reasoning for power operation risk control, and the update of the case base and the generation of decision-making plans. In the stage of constructing the ontology model for power operation risk control, according to the power operation risk control record report, important concepts and terms in the record are extracted, classes and class hierarchy relationships are defined based on ontology technology, and attribute characteristics are defined to construct a power accident ontology model, and the case knowledge is structurally represented and stored as a case base. In the stage of case-based reasoning for power operation risk control, the attribute values of the input risk control target case are calculated, and the attribute similarity is calculated with the cases in the case base. The entropy weight method is used to calculate the attribute weights, and then the global similarity calculation based on the grey algorithm is further adopted to match the most similar historical case from the case base. In the stage of case base update and decision-making plan generation, based on the matched similar cases, a targeted auxiliary control plan is generated, and the new cases are stored to automatically update the case base.

[0078] The present invention applies ontology and case-based reasoning technologies to the risk control of electric power operations. The ontology technology constructs a knowledge ontology hierarchy for electric power operation risk control by structurally representing the case knowledge of electric power operation reports, clarifying concepts and their relationships, providing a solid knowledge foundation for case-based reasoning. The case-based reasoning technology realizes the rapid retrieval and reuse of case knowledge by calculating the similarity between cases, thereby assisting decision-makers in formulating more scientific and reasonable accident response plans and providing strong support for the risk control of accidents at electric power construction sites.

[0079] The following takes a specific embodiment of a power grid operation personal accident report for a more detailed explanation and description. It should be understood that the described embodiments are exemplary embodiments of the present invention, rather than all embodiments, and are intended to explain the present invention, rather than being construed as a limitation of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present invention.

[0080] As Figure 1 shown, the specific steps of the electric power operation risk control method of the present invention are as follows:

[0081] S1. Analyze the records of electric power operation personal accident reports and construct a knowledge base of the ontology of electric power operation personal accidents.

[0082] According to the investigation report of electric power operation personal accidents, specific accident cases include accident basic information, accident process information, and accident analysis information. Among them, the accident basic information includes accident number, accident name, occurrence time, occurrence location, accident unit, accident type, economic loss, and personnel casualties, etc.; the accident process information is a detailed description of the accident occurrence process to clarify the accident responsibility; the accident analysis information is the accident cause and lessons learned obtained through the analysis and induction of industry experts. The ontology knowledge base is the structured storage of the basic information, process information, and analysis information of historical accidents, facilitating subsequent case-based reasoning and providing decision-making suggestions.

[0083] S2. Convert the accident cases into data structures recognizable by a computer, including accident basic information, process information, and analysis information, to realize the construction of a risk control knowledge ontology model (electric power accident ontology model).

[0084] The ontology model of power accidents includes a concept set C_Case, a relation set R_Case, an instance set I_Case, and an axiom set A_Case. The concepts involved in C_Case include "accident", "accident causation factors", "accident level", and "accident type". Taking the concept of "accident" as an example, Table 1 defines the basic attributes and attribute types of this concept; the relations involved in R_Case include the part-of relationship between part and whole, the kind-of inheritance between concepts, the instance-of relationship between instance and concept, and the attribute-of relationship between attribute and concept. Table 2 illustrates the meaning of the relations by examples; I_Case represents the instance set of concepts, which refers to the instantiation set of the concepts defined by C_Case and describes the specific data of actual power accidents. For example, the detailed information of a certain accident, including the specific values of concepts such as accident number, accident name, occurrence time, occurrence location, accident unit, number of deaths, number of injuries, and economic losses. A_Case represents the axiom set, and the axiom set defines the rules that should be followed in this ontology model, including the value-taking rules of some categorical variables and the normalization requirements for the value-taking of some concepts. The rules are determined based on the experience of power safety accident handling and the characteristics of historical accident data. Table 4 lists the axiom set in the ontology model of power accidents.

[0085] Table 1 Definition of the "Accident" Concept in the Ontology of Personal Injury Accidents in Power Operations

[0086]

[0087]

[0088] Table 2 Definition of the Basic Relations of Concepts in the Ontology of Personal Injury Accidents in Power Operations

[0089] Serial number Relationship name Domain Range 1 kind-of Human factor Accident-causing factor 2 instance-of Electric shock Accident type 3 attribute-of Occurrence time Accident

[0090] Table 3 Case Feature Attributes

[0091]

[0092] Table 4 List of Rules in the Ontology of Personal Injury Accidents in Power Operations

[0093]

[0094]

[0095] It should be understood that in the ontology model, the definitions of concepts and relationships usually include the following three basic elements: name, domain, and range. The attributes in Table 1 describe the general characteristics of a certain concept, apply to all instances under this concept, and are the basic elements used to define concepts in model design. The attributes in Table 3 are the attributes for specific accident instances, recording the detailed information of specific accidents, and are the elements used to describe the actual situation in model application. In an instance, the attribute definition usually includes some basic elements such as name, code, type, value, etc. However, whether it includes the "concept" in the concept set and the determination method of attribute indicators can vary according to specific model design requirements. The attribute does not necessarily directly include the "concept" in the concept set. Usually, the concepts in the concept set are used to define the framework of attributes and relationships, while in specific instances, the attribute mainly records the specific values of the concept. For example, "accident type" as a concept attribute may be concretized into actual values such as "fall from height" or "electric shock" in an instance. The determination of attribute indicators is based on multiple factors such as model requirements, concept definitions, industry standards, data sources, and application scenarios.

[0096] According to the existing sample information, taking accident case knowledge as an example, on August 27, 2014, an electric shock accident occurred in a certain engineering company. Case details: On August 27, 2014, when a certain engineering company was carrying out the construction of the 10kV substation transformation project of Hainan Sanya Power Supply Bureau in Haitangwan Town, Sanya City, Hainan Province, an electric shock accident occurred to the personnel of the outsourced unit, resulting in 1 death. The extraction of accident case knowledge is as Figure 2 shown.

[0097] S3. Case similarity calculation

[0098] Calculate the similarity between the sudden accident and the cases in the case library. Regarding the sudden accident as the target case, calculate the local similarity of the case, including attribute similarity and structure similarity; calculate the concept similarity of the case, including tree distance similarity and WUP similarity, and combine the grey algorithm to weight and calculate the concept similarity; and combine the global similarity calculation model to improve the accuracy of case matching. Refer to Figure 3 , which specifically includes the following content:

[0099] S3-1 Local similarity calculation, including the calculation of attribute similarity and structure similarity.

[0100] 1) Attribute similarity. Different calculation methods are adopted for numerical and symbolic attributes of attribute similarity to ensure the accuracy of similarity. The formula is as follows:

[0101] I. Numerical attributes

[0102]

[0103] In the formula, case X* is the target case, X i represents the i-th case in the case base, max j represents the maximum value of the j-th characteristic attribute, min j represents the minimum value of the j-th characteristic attribute, represents the value of the j-th characteristic attribute of the target case, x ij represents the value of the j-th characteristic attribute of the i-th case.

[0104] II. Symbolic Attributes

[0105] Symbolic attributes belong to the enumeration type and can be specifically divided into two types: ordered and unordered. Among the characteristic attributes of power operation personal accident cases, the accident type belongs to the unordered enumeration type variable; the accident level belongs to the ordered enumeration type variable. The similarity calculation methods for the above two types of attributes are as follows.

[0106] i. When the characteristic attribute is an unordered enumeration type variable,

[0107]

[0108] ii. When the characteristic attribute is an ordered enumeration type variable,

[0109]

[0110] where count j represents the number of values of the j-th characteristic attribute.

[0111] 2) Structural similarity. Structural similarity considers information loss or improper recording and solves the problem of missing attribute values through a specific calculation method to improve the accuracy of global similarity. Due to information loss or improper recording in the operation report, directly summing the weighted characteristic attributes of the cases will result in a low global similarity. Therefore, before calculating the global similarity, first calculate the structural similarity of the cases to solve the problem of inaccurate global similarity caused by missing some attribute values. Denote the structural similarity between the target case X * and the case X i as Sim stru (X * ,X i ), and the calculation method is as follows.

[0112]

[0113] where W Q∩C is the number of elements in the intersection of the conceptual attributes of the target case and the i-th case, W Q∪C is the number of elements in the union of the conceptual attributes of the target case and the i-th case, w i is the weight of the i-th attribute in the intersection, w kis the weight of the k-th attribute in the union set, and m is the value of W Q∩C 's value, and l is the value of W Q∪C 's value.

[0114] S3-2 Concept Similarity Calculation Based on Grey Algorithm

[0115] For concept similarity calculation, especially for the comparison of accident causation concepts, a "concept tree" similarity calculation method based on semantic information is adopted. The similarity between two cases in terms of causation concepts is calculated through the definitions of tree distance similarity, WUP similarity, and concept similarity.

[0116] 1) Tree Distance Similarity

[0117] In the concept tree, the tree similarity Sim 1 of concepts c 2 and c t (c 1 , c 2 ) is calculated as follows:

[0118]

[0119] In the formula, f(c 1 ) and f(c 2 ) respectively represent the concept levels of c 1 and c 2 ; d(c 1 , c 2 ) is the concept distance between c 1 and c 2 ; a > 0 is a tuning parameter, and Sim t (c 1 , c 2 ) = 1 indicates an equivalence relationship between concepts c 1 and c 2 .

[0120] 2) WUP Similarity

[0121] The WUP similarity is the similarity degree between upper-level concepts, reflecting the indirect similarity between concepts, denoted as Sim s (c 1 , c 2 ), and is calculated according to the calculation formula (5) of tree distance similarity. WUP focuses on the similarity degree of the upper-level concepts of concepts, that is, the similarity of the upper-level concepts calculated by the tree similarity algorithm.

[0122] 3) The concept similarity between two non-equivalent concepts c 1 and c 2 is denoted as Sim(c 1 , c 2 ) and is calculated as follows

[0123]

[0124] Let the target case be X * contain m cause concepts {p 11 , p 12 ,..., p 1m}, and X i contain n cause concepts. Denote the cause concept similarity between case X * and X i as Sim o (X * , X i ). The calculation steps are as follows:

[0125] I. Calculate the similarity matrix between X * and X i according to the methods in steps 1) to 3) above:

[0126] SimM ij = Sim c (c i , c j ) (7)

[0127] II. Find the maximum similarity sequence. Take the maximum value λ jz in the similarity matrix, and delete the row and column where this maximum value is located to obtain matrix S'. Repeat this step until the matrix is empty. Denote each obtained maximum value λ iz as where t 0 = min(m, n). Add each obtained maximum value to the maximum similarity sequence , and calculate the existing factor weight w i using the gray algorithm. Here, both the existing factor weight and the index weight express the importance degree of the corresponding attribute or concept. Only the index weight (i.e., the attribute weight) is calculated by the entropy weight method, while the existing factor weight (i.e., the concept weight) is calculated by the gray algorithm. The calculation steps are as follows:

[0128] i. Perform mean value processing on the sequence. The formula is as follows:

[0129]

[0130] ii. Solve the absolute value sequence of the mean value sequence:

[0131] Δ(k) = |SimM ij ′(k) - Sim max ′(k)|, k = 1, 2,..., t 0 (10)

[0132] iii. Find the maximum and minimum values in the above absolute value sequence, denoted as F and f respectively;

[0133] iv. Calculate the correlation coefficient and correlation degree at each time point of the sequence:

[0134]

[0135] Here, the correlation degree w i has the same meaning as the aforementioned existing factor weights (concept weights), and the expression of the correlation degree is used as the weight of the concept.

[0136] III. Combine the structural similarity Sim stru (X * , X i )(The calculation formula is shown in Formula 4) to calculate the cause concept similarity, where B is an adjustment parameter:

[0137]

[0138] S3-3 Global similarity calculation

[0139] Through local similarity, concept similarity, and feature attribute weights, the global similarity between the target case and the alternative cases can be further solved.

[0140]

[0141] w j represents the weight of the jth attribute, and Sim j (X * , X i ) represents the local similarity of the sample case X i and the target case X * in the jth attribute, and Sim stru (X * , X i ) represents the case structure similarity, and Sim o (X * , X i ) represents the cause concept similarity between the sample case X i and the target case X * .

[0142] Among them, the calculation method of the weight parameter w j is as follows: First, construct the membership evaluation matrix R, then standardize the values of each index to convert them into the same dimension, then calculate the proportion of the characteristic attribute values of each case in the total characteristic attribute values of the case library for this index, then calculate the information entropy of the characteristic attribute values of each index, and finally calculate the weights of the characteristic attribute values of each index.

[0143] 1) Suppose the case base contains a total of n cases and m characteristic attribute values, forming a membership evaluation matrix R:

[0144]

[0145] 2) First, it is necessary to standardize the values of each index and convert them to the same dimension. There are two methods of standardization: when the index value is better when it is larger, that is, a positive index, its standardization formula is When the index value is better when it is smaller, that is, a negative index, its standardization formula is In the present invention, the data are all negative indexes, and r ij is the membership degree of the evaluation index.

[0146] 3) Calculate the proportion p ij of the j-th characteristic attribute value of the i-th case to the total characteristic attribute value of the index in the case base, and there is:

[0147]

[0148] 4) Calculate the entropy value e j of the j-th index characteristic attribute value, and there is:

[0149]

[0150] 5) Calculate the information entropy d j of the j-th index characteristic attribute value, and there is

[0151] d j = 1 - e j (18)

[0152] 6) Calculate the weight w j of the j-th index characteristic attribute value, and there is

[0153]

[0154] S4. Based on the calculation results, perform case matching to provide intelligent support for decision-makers.

[0155] According to the calculation results of the global case similarity, generate an auxiliary control and management plan for the new case. Realize the storage of the new case and complete the automatic update of the case base to adapt to the new accident scenario. If no suitable alternative cases can be screened out through the attached Figure 1 analysis, it is necessary to make necessary adjustments to the existing case data. This includes re-examining and revising the criteria and rules for case selection to construct a more accurate case base. After the correction is completed, recalculate the global case similarity using the specified formula.

[0156] Aiming at the problem of difficult emergency decision-making brought about by the complex on-site environment in power operation risk control, the present invention proposes a calculation framework and application method for similarity analysis and case matching of power operation report cases based on similarity matching of the grey algorithm through an ontology, case reasoning, and a method based on the similarity of the grey algorithm. The present invention is illustrated by combining specific case fields on this basis. The present invention identifies and incorporates those cases that make significant contributions to improving the quality of decision-making according to established case learning rules. By continuously enriching the content of the case library, the timeliness and reliability of future operation risk control are improved.

[0157] Embodiment 2

[0158] This embodiment provides a power operation risk control system, which includes:

[0159] An ontology knowledge base construction module, which is used to analyze the records of power operation personal accident reports and form a power operation personal accident ontology knowledge base;

[0160] An ontology model construction module, which is used to define a concept set, a relationship set, an instance set, and an axiom set based on the ontology model according to the power operation personal accident ontology knowledge base, and obtain a power operation personal accident ontology model according to the concept set, the relationship set, the instance set, and the axiom set;

[0161] A case reasoning module, which is used to obtain a target case, calculate the similarity between the target case and each case in the case library based on the grey algorithm, and obtain the case in the case library that is closest to the target case as a similar case according to the similarity;

[0162] A risk control assistance module, which is used to generate an auxiliary control plan for the accident associated with the target case according to the retrieved and matched similar cases.

[0163] Among them, the concept set includes accidents, accident-causing factors, accident levels, and accident types, and each concept includes basic attributes and attribute types; the relationship set includes the relationships between parts and the whole, inheritance between concepts, instances and concepts, and attributes and concepts; the instance set is an instance set of concepts, which is used to describe the specific data of actual power accidents; the axiom set includes the value-taking rules of categorical variables and the normalization requirements for concept value-taking, and the rules are determined based on power safety accident handling experience and historical accident data characteristics.

[0164] Furthermore, the case reasoning module includes:

[0165] A local similarity calculation unit, which is used to calculate the local similarity between a target case and historical cases, including attribute similarity and structural similarity. Different calculation methods are adopted for numerical and symbolic attributes for the attribute similarity, and the structural similarity is calculated based on the relationship between the intersection and union of the concept attributes of the target case and each case in the case library;

[0166] A cause concept similarity calculation unit, which is used to calculate the similarity between two cases in terms of cause concepts by using the grey algorithm and through the definitions of tree distance similarity, WUP similarity and concept similarity;

[0167] A global similarity calculation unit, which is used to sum the weighted attribute similarities according to different weights of attribute indicators, and combine the structural similarity and the similarity in cause concepts to obtain the global similarity;

[0168] A similar case determination unit, which is used to obtain the case with the highest global similarity in the case library as the similar case of the target case.

[0169] Furthermore, different calculation methods are adopted for numerical and symbolic attributes for the attribute similarity, including:

[0170] For numerical attributes, the following formula is used for calculation:

[0171]

[0172] In the formula, case X * is the target case, and X i represents the i-th case in the case library, and max j represents the maximum value of the j-th characteristic attribute; min j represents the minimum value of the j-th characteristic attribute; represents the value of the j-th characteristic attribute of the target case, and x ij represents the value of the j-th characteristic attribute of the i-th case;

[0173] For symbolic attributes: when the characteristic attribute is an unordered enumerated variable, the following formula is used for calculation:

[0174]

[0175] When the characteristic attribute is an ordered enumerated variable, the following formula is used for calculation:

[0176]

[0177] Among them, count j represents the number of values of the j-th characteristic attribute.

[0178] Furthermore, the structural similarity is calculated according to the following formula:

[0179]

[0180] Where W Q∩C is the number of elements in the intersection of the target case and the conceptual attributes of the i-th case, and W Q∪C is the number of elements in the union of the target case and the conceptual attributes of the i-th case, and w i is the weight of the i-th attribute in the intersection, and w k is the weight of the k-th attribute in the union. m is the value of W Q∩C and l is the value of W Q∪C .

[0181] Furthermore, using the grey algorithm, the similarity between two cases in terms of causal concepts is calculated through the definitions of tree distance similarity, WUP similarity, and concept similarity, including:

[0182] Let the target case X * contain m causal concepts {p 11 , p 12 ,..., p 1m}, and the i-th case X i in the case base contain n causal concepts. Denote the causal concept similarity between cases X * and X i as Sim o (X * , X i ). The calculation steps are as follows:

[0183] I. Denote the tree similarity between concepts c t and c 1 as Sim 2 (c 1 , c 2 ), and the WUP similarity between the upper-level concepts of c s and c 1 as Sim 2 (c 1 , c 2 ). Then the concept similarity Sim 1 (c 2 , c c ) between concepts c 1 and c 2 is calculated by the formula ; According to the above calculation method, the similarity matrix between X * and X i is obtained: SimM ij = Sim c (c i , c j );

[0184] II. Find the sequence with the maximum similarity, and take the maximum value λ in the similarity matrix ij , and delete the row and column where the maximum value is located to obtain the matrix S′. Repeat this step until the matrix is empty; record the maximum value λ obtained each time ij as where t 0 = min(m, n); add the maximum value obtained each time to the sequence of the maximum similarity , and calculate the weight w of the existing factors using the grey algorithm i ;

[0185] III. Calculate the similarity of cause concepts according to the following formula: where B is a regulation parameter, and Sim stru (X * , X i ) is the structural similarity between two cases.

[0186] Furthermore, the global similarity is calculated as follows:

[0187]

[0188] where w j represents the weight of the j-th attribute, Sim j (X * , X i ) represents the attribute similarity of case X i and the target case X * in the j-th attribute, Sim stru (X * , X i ) represents the structural similarity between case X i and the target case X * , and Sim o (X * , X i ) represents the similarity of cause concepts between case X i and the target case X * .

[0189] It should be understood that the power operation risk control system based on grey algorithm similarity matching in the embodiments of the present invention can implement all the technical solutions in the above method embodiments. The functions of its various functional modules can be specifically implemented according to the methods in the above method embodiments. For the specific calculation methods of the attribute similarity, structural similarity, concept similarity, and global similarity, the specific implementation process can refer to the relevant descriptions in the above embodiments, which will not be elaborated here.

[0190] Example 3

[0191] This embodiment provides a computer device, which includes one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and when the programs are executed by the processors, the steps of the power operation risk control method based on grey algorithm similarity matching as described above are implemented.

[0192] Embodiment 4

[0193] This embodiment provides a computer storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the power operation risk control method based on grey algorithm similarity matching as described above are implemented.

[0194] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a device (system), a computer device, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code.

[0195] The present invention is described with reference to the flowcharts of the methods according to the embodiments of the present invention. It should be understood that each process in the flowchart and the combination of the processes in the flowchart can be implemented by computer program instructions. These computer program instructions can be provided to the processors of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processors of the computer or other programmable data processing devices generate a device for implementing the functions specified in one Figure 1 process or multiple processes.

[0196] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the functions specified in one Figure 1 process or multiple processes.

[0197] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one Figure 1 process or multiple processes.

Claims

1. A method for risk management and control of power operations, characterized in that: The method comprises the following steps: Analyze the reports and records of personal accidents in power operations to form a knowledge base of personal accidents in power operations; According to the ontology knowledge base of personal accidents in power operation, a concept set, a relationship set, an instance set, and an axiom set are defined based on the ontology model, and an ontology model of personal accidents in power operation is obtained according to the concept set, the relationship set, the instance set, and the axiom set; Obtain the target case, calculate the similarity between the target case and each case in the case library based on the grey algorithm, and obtain the case in the case library that is closest to the target case as a similar case based on the similarity; Based on the similar cases obtained through retrieval and matching, an auxiliary control plan for the accident associated with the target case is generated.

2. The method according to claim 1, characterized in that The concept set includes accidents, accident causes, accident levels, and accident types. Each concept contains basic attributes and attribute types; the relationship set includes the relationship between part and whole, inheritance between concepts, instances and concepts, and attributes and concepts; the instance set is an instance set of concepts, which is used to describe the specific data of actual power accidents; the axiom set includes the value-taking rules of categorical variables and the standardization requirements for concept values. The rules are determined based on the experience of handling power safety accidents and the characteristics of historical accident data.

3. The method according to claim 1, characterized in that The similarity between the target case and each case in the case library is calculated based on the grey algorithm, and the case in the case library that is closest to the target case is obtained as a similar case according to the similarity, including: Calculate the local similarity between the target case and the historical cases, including attribute similarity and structural similarity. The attribute similarity uses different calculation methods for numerical and symbolic attributes. The structural similarity is calculated based on the relationship between the intersection and union of the concept attributes of the target case and each case in the case library. The grey algorithm is used to calculate the similarity of the two cases in the causal concept through the definitions of tree distance similarity, WUP similarity and concept similarity; According to the different weights of attribute indicators, the similarities of each attribute are weighted and summed, and the global similarity is obtained by combining the structural similarity and the similarity of the causal concept; The case with the highest global similarity in the case library is obtained as a similar case to the target case.

4. The method according to claim 3, characterized in that The attribute similarity adopts different calculation methods for numerical and symbolic attributes, including: For numerical attributes, the following formula is used for calculation: In the formula, case X * For the target case, X i represents the i-th case in the case library, max j Indicates the maximum value of the j-th feature attribute; min j Represents the minimum value of the jth feature attribute; represents the value of the j-th feature attribute of the target case, x ij represents the value of the jth feature attribute of the i-th case; For symbolic attributes: When the characteristic attribute is an unordered enumeration variable, the following formula is used for calculation: When the characteristic attribute is an ordered enumeration variable, the following formula is used for calculation: Among them, count j Indicates the number of values ​​of the j-th feature attribute.

5. The method according to claim 3, characterized in that: The structural similarity is calculated according to the following formula: Where W Q∩C is the number of elements in the intersection of the target case and the i-th case concept attributes, W Q∪C is the number of elements in the union of the target case and the i-th case concept attributes, w i is the weight of the i-th attribute in the intersection, w k is the weight of the kth attribute in the union, and m is W Q∩C The value of l is W Q∪C The value of .

6. The method according to claim 3, characterized in that The grey algorithm is used to calculate the similarity of the causal concepts of the two cases through the definitions of tree distance similarity, WUP similarity and concept similarity, including: Set Target Case X * Contains m cause concepts {p 11 ,p 12 ,...,p 1m }, the i-th case X in the case library i Contains n cause concepts, case X * and X i The similarity of the cause concept between o (X * ,X i ), the calculation steps are as follows: I. Sim t (c1,c2) ​​represents the tree similarity between concepts c1 and c2, expressed as Sim s (c1,c2) ​​represents the WUP similarity between the upper-level concepts of c1 and c2, then the concept similarity Sim between concepts c1 and c2 c (c1,c2) ​​by the formula Calculated; According to the above calculation method, X * and X i Similarity matrix between: SimM ij =Sim c (c i ,c j ); II. Find the maximum similarity sequence and take the maximum value λ in the similarity matrix ij , and delete the row and column where the maximum value is located to obtain the matrix S′, and repeat this step until the matrix is ​​empty; record the maximum value λ obtained each time ij for Where t0 = min(m,n); add the maximum value obtained each time to the maximum similarity sequence In the above example, the grey algorithm is used to calculate the existing factor weight w i ; III..Calculate the cause concept similarity according to the following formula: Where B is a tuning parameter, Sim stru (X * ,X i ) is the structural similarity between two cases.

7. The method according to claim 3, characterized in that The global similarity is calculated as follows: Among them, w j represents the weight of the jth attribute, Sim j (X * ,X i ) represents case X i With target case X * The attribute similarity of the jth attribute, Sim stru (X * ,X i ) represents case X i With target case X * The structural similarity between o (X * ,X i ) represents case X i With target case X * The similarity of the cause concepts between them.

8. The method according to claim 1, characterized in that: The method further includes: after each auxiliary management and control plan is generated, it is stored as a new case in the case library to achieve automatic updating of the case library.

9. A power operation risk management and control system, characterized in that: The system comprises: The ontology knowledge base construction module is used to analyze the report records of personal accidents in power operations and form an ontology knowledge base of personal accidents in power operations; An ontology model building module is used to define a concept set, a relationship set, an instance set, and an axiom set based on an ontology model according to an ontology knowledge base of personal accidents in power operation, and obtain an ontology model of personal accidents in power operation according to the concept set, the relationship set, the instance set, and the axiom set; The case reasoning module is used to obtain the target case, calculate the similarity between the target case and each case in the case library based on the grey algorithm, and obtain the case in the case library that is closest to the target case as a similar case based on the similarity; The risk management and control auxiliary module is used to generate an auxiliary management and control plan for the accident associated with the target case based on similar cases obtained through retrieval and matching.

10. The system according to claim 9, characterized in that The concept set includes accidents, accident causes, accident levels, and accident types. Each concept contains basic attributes and attribute types; the relationship set includes the relationship between part and whole, inheritance between concepts, instances and concepts, and attributes and concepts; the instance set is an instance set of concepts, which is used to describe the specific data of actual power accidents; the axiom set includes the value-taking rules of categorical variables and the standardization requirements for concept values. The rules are determined based on the experience of handling power safety accidents and the characteristics of historical accident data.

11. The system according to claim 9, characterized in that The case-based reasoning module includes: A local similarity calculation unit is used to calculate the local similarity between the target case and the historical case, including attribute similarity and structural similarity. The attribute similarity adopts different calculation methods for numerical and symbolic attributes. The structural similarity is calculated based on the relationship between the intersection and union of the concept attributes of the target case and each case in the case library. The cause concept similarity calculation unit is used to calculate the similarity of the cause concepts of two cases by using the grey algorithm through the definitions of tree distance similarity, WUP similarity and concept similarity; The global similarity calculation unit is used to perform weighted summation of the similarities of each attribute according to the different weights of the attribute indicators, and combine the structural similarity and the similarity of the causal concept to obtain the global similarity; The similar case determination unit is used to obtain the case with the highest global similarity in the case library as a similar case to the target case.

12. The system according to claim 11, characterized in that The attribute similarity adopts different calculation methods for numerical and symbolic attributes, including: For numerical attributes, the following formula is used for calculation: In the formula, case X * For the target case, X i represents the i-th case in the case library, max j Indicates the maximum value of the j-th feature attribute; min j Represents the minimum value of the jth feature attribute; represents the value of the j-th feature attribute of the target case, x ij represents the value of the jth feature attribute of the i-th case; For symbolic attributes: When the characteristic attribute is an unordered enumeration variable, the following formula is used for calculation: When the characteristic attribute is an ordered enumeration variable, the following formula is used for calculation: Among them, count j Indicates the number of values ​​of the j-th feature attribute.

13. The system according to claim 11, characterized in that The structural similarity is calculated according to the following formula: Where W Q∩C is the number of elements in the intersection of the target case and the i-th case concept attributes, W Q∪C is the number of elements in the union of the target case and the i-th case concept attributes, w i is the weight of the i-th attribute in the intersection, w k is the weight of the kth attribute in the union, and m is W Q∩C The value of l is W Q∪C The value of .

14. The system according to claim 11, characterized in that The grey algorithm is used to calculate the similarity of the causal concepts of the two cases through the definitions of tree distance similarity, WUP similarity and concept similarity, including: Set Target Case X * Contains m cause concepts {p 11 ,p 12 ,...,p 1m }, the i-th case X in the case library i Contains n cause concepts, case X * and X i The similarity of the cause concept between o (X * ,X i ), the calculation steps are as follows: I. Sim t (c1,c2) ​​represents the tree similarity between concepts c1 and c2, expressed as Sim s (c1,c2) ​​represents the WUP similarity between the upper-level concepts of c1 and c2, then the concept similarity Sim between concepts c1 and c2 c (c1,c2) ​​by the formula Calculated; According to the above calculation method, X * and X i Similarity matrix between: SimM ij =Sim c (c i ,c j ); II. Find the maximum similarity sequence and take the maximum value λ in the similarity matrix ij , and delete the row and column where the maximum value is located to obtain the matrix S′, and repeat this step until the matrix is ​​empty; record the maximum value λ obtained each time ij for Where t0 = min(m,n); add the maximum value obtained each time to the maximum similarity sequence In the above example, the grey algorithm is used to calculate the existing factor weight w i ; III..Calculate the cause concept similarity according to the following formula: Where B is a tuning parameter, Sim stru (X * ,X i ) is the structural similarity between two cases.

15. The system according to claim 12, characterized in that The global similarity is calculated as follows: Among them, w j represents the weight of the jth attribute, Sim j (X * ,X i ) represents case X i With target case X * The attribute similarity of the jth attribute, Sim stru (X * ,X i ) represents case X i With target case X * The structural similarity between o (X * ,X i ) represents case X i With target case X * The similarity of the cause concepts between them.

16. A computer device, characterized in that: The device includes one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and when the programs are executed by the processors, the steps of the power operation risk management method based on gray algorithm similarity matching as described in any one of claims 1 to 8 are implemented.

17. A computer storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the power operation risk management method based on grey algorithm similarity matching as described in any one of claims 1 to 8 are implemented.

Images