Risk assessment method for power transmission and transformation construction based on variable weight mahalanobis distance matter-element extension

Abstract

The application provides a power transmission and transformation construction risk evaluation method based on variable weight Mahalanobis distance matter element extension, first constructs a power transmission and transformation engineering construction safety risk evaluation index system, adopts a BWM method to calculate a subjective weight, adopts a CRITIC method to calculate an objective weight, obtains a basic constant weight through a game theory combination weighting method; secondly, constructs a classical domain matter element matrix, a section domain matter element matrix and a matter element to be evaluated, normalizes the to-be-evaluated data to obtain a real-time state numerical vector, and extracts a risk grade characteristic center vector; thirdly, introduces a penalty type state variable weight function to construct a state variable weight vector, and obtains a final variable weight weight vector; finally, calculates a variable weight Mahalanobis distance between the real-time state numerical vector and the characteristic center vector, obtains a risk grade closeness degree, and determines a final risk grade. The application eliminates linear correlation interference between the evaluation indexes by introducing the Mahalanobis distance, simultaneously overcomes a neutralization effect of a traditional constant weight evaluation by using a penalty type variable weight mechanism, and significantly improves the accuracy and sensitivity of the power transmission and transformation engineering construction site safety risk grade determination.

Description

Technical Field

[0001] This invention relates to the field of power transmission and transformation engineering construction technology, specifically to a method for assessing the construction risks of power transmission and transformation based on the extension of matter-element with variable weight Mahalanobis distance. Background Technology

[0002] Power transmission and transformation engineering construction encompasses civil engineering, electrical engineering, overhead line engineering, and cable line engineering, involving multiple stages such as high-altitude operations, large machinery operation, and electrical equipment installation. Especially in the construction of substation expansion and renovation projects and transmission line projects, the construction is characterized by proximity to energized bodies, relocation of work sites, crossing of energized lines, and crossing of highways, railways, rivers, and lakes. The construction environment is extremely complex, with significant safety risks, making it a highly comprehensive and systematic project. Once a safety accident occurs, it will not only cause huge casualties and property losses but also seriously affect the safe and stable operation of the power grid. Therefore, scientifically and accurately identifying and evaluating the safety risk level of the construction site is of great significance for improving the safety management level of power transmission and transformation projects.

[0003] Currently, the construction site environment for power transmission and transformation projects is highly variable, with complex and intertwined relationships between various risk factors. Furthermore, the loss of control over a single critical link can often trigger an overall safety accident, exhibiting significant localized degradation characteristics. However, existing quantitative tools such as the Analytic Hierarchy Process (AHP) and fuzzy comprehensive evaluation methods have obvious limitations in addressing these engineering characteristics. On the one hand, traditional methods typically rely on the assumption of independence between evaluation indicators, ignoring the strong correlations that objectively exist between various risk factors at the construction site. This calculation method, which fails to eliminate information overlap between indicators, easily leads to biased evaluation results. On the other hand, existing methods often employ a fixed-weight constant-weight evaluation model. When a critical indicator deteriorates drastically, its risk impact can easily be neutralized by other indicators in good condition, resulting in evaluation results that fail to sensitively reflect significant localized safety hazards at the construction site and cannot meet the actual needs for accurate risk identification in high-risk construction environments. Therefore, there is an urgent need to research a power transmission and transformation construction risk assessment method based on variable-weight Mahalanobis distance matter-element extension that can effectively eliminate interference from correlations between indicators and possess variable-weight adjustment capabilities. Summary of the Invention

[0004] In view of this, the present invention provides a method for risk assessment of power transmission and transformation construction based on variable-weight Mahalanobis distance matter-element extension. This method aims to solve the problems of evaluation distortion caused by the difficulty in eliminating interference from correlations between evaluation indicators and the inability of constant-weight evaluation models to sensitively reflect major local safety hazards in existing evaluation techniques. By introducing an inverse covariance matrix of evaluation indicators to eliminate interference from indicator correlations and combining it with a penalized variable-weight mechanism for accurate evaluation, the method improves the accuracy and sensitivity of determining the safety risk level of power transmission and transformation construction. To achieve the above objectives, the present invention adopts the following technical solution: The method for risk assessment of power transmission and transformation construction based on the extension of the matter-element with variable weight Markov distance includes the following steps: Step S1: Introduce a framework for analyzing and classifying human factors in construction to build a preliminary indicator system for evaluating the construction safety risks of power transmission and transformation projects. Through data collection, expert consultation, questionnaires, and data analysis, positive verification, reverse verification, and actual project feedback verification are conducted to select multiple evaluation indicators and form the final indicator system for evaluating the construction safety risks of power transmission and transformation projects. Step S2: Calculate the subjective and objective weights of each evaluation index using the BWM method and the CRITIC method respectively, and then fuse the subjective and objective weights using the game theory combination weighting method to obtain the basic constant weight vector of the evaluation index. Step S3: Construct the classical domain matter-element matrix, the section domain matter-element matrix, and the matter-element to be evaluated for the risk assessment sample. Normalize the measured data to be evaluated based on the section domain matter-element boundary to determine the real-time state value vector of the matter-element to be evaluated. Extract the feature center vector of each risk level from the classical domain matter-element matrix based on the same normalization rule. Step S4: Construct a penalized state-weighted function. Based on the real-time state value vector of the object to be evaluated, and combined with the preset penalty threshold and penalty intensity factor, construct the state-weighted vector of the object to be evaluated. Then, use the state-weighted vector to adaptively correct the basic constant weight vector and calculate the final weighted vector of the object to be evaluated. Step S5: Construct the covariance matrix and its inverse matrix of the evaluation index, convert the final variable weight vector of the object to be evaluated into a diagonal matrix, and then use the covariance inverse matrix and the diagonal matrix to calculate the variable weight Mahalanobis distance between the real-time state value vector of the object to be evaluated and the feature center vector of each risk level. Step S6: Based on the variable weight Mahalanobis distance, calculate the proximity of the object to be evaluated with respect to each risk level, and determine the final risk level of the construction safety of the power transmission and transformation project according to the principle of maximum proximity.

[0005] In step S1, the indicators in the final power transmission and transformation project construction safety risk assessment index system are divided into two levels; The primary indicators include the impact of corporate organization, safety supervision, on-site operation-related factors, and construction personnel-related factors; The secondary indicators are as follows: organizational impact includes organizational structure and responsibilities, safety production investment, and safety management procedures; safety supervision includes risk monitoring and early warning, supervision and management violations, and work plan arrangements; on-site operation related factors include basic construction conditions, construction technology measures, construction machinery, construction natural environment, and construction social environment; and construction personnel related factors include violations of regulations, skill errors, intuition and decision-making errors, and personnel quality.

[0006] In step S2, the specific process of calculating the subjective weights of the evaluation indicators using the BWM method is as follows: Determine the optimal indicator in the risk assessment indicator system and worst indicators ; Determine the importance preference of the optimal indicator to other indicators, and construct the optimal indicator comparison vector. ,in Indicates the optimal index relative to the first The importance preference of each indicator; Determine the importance preference of other indicators to the worst indicator, and construct a comparison vector for the worst indicator. ,in Indicates the first The importance preference of each indicator relative to the worst indicator; Construct the following linear BWM optimization model:

[0007] In the formula, The subjective weight of the optimal indicator; The subjective weight of the worst-case indicator; For the first Subjective weighting of each indicator; For consistency parameters; Therefore, the optimal consistency parameters can be obtained by solving the model. Subjective weights of evaluation indicators and subjective weight vector .

[0008] To verify the reliability of the expert scoring logic, a consistency ratio is introduced. Perform a consistency check: Based on the maximum importance preference value used in the comparison vector, the corresponding consistency index is obtained by looking up the BWM Consistency Index (CI) mapping table. ; According to the formula Calculate the consistency ratio. If the consistency ratio is... If the value is less than the preset validity threshold, the consistency check is considered passed; otherwise, the comparison vector needs to be re-evaluated. and The model is revised to eliminate conflicts in the subjective logic of the experts until it passes the test, thereby ensuring the robustness of the evaluation model.

[0009] In step S2, the specific process of calculating the objective weights of the evaluation indicators using the CRITIC method is as follows: Build includes One sample Initial data matrix of each evaluation indicator Then, normalization is performed to obtain a standardized matrix. ; Calculate the first Standard deviation of each indicator and the The first indicator and the first Pearson correlation coefficient among the indicators ; Calculate the first according to the following formula. Conflicts among indicators and amount of information :

[0010]

[0011]

[0012] Calculate the objective weights of each evaluation indicator. To obtain the objective weight vector .

[0013] In step 2, the specific process of calculating the basic constant weight vector using the game theory combinatorial weighting method is as follows: Construct a linear combination vector of subjective and objective weights ,in and These are the combination coefficients of subjective weights and objective weights, respectively; and These are the subjective weight vector and the objective weight vector, respectively. Based on the principle of minimizing deviation, a method for solving the optimal combination coefficients is constructed. Linear equation model:

[0014] Solving the system of linear equations yields and The final combination coefficients are obtained by normalizing them according to the following formula. and :

[0015] According to the formula The basic constant weight vectors of each evaluation index are calculated. .

[0016] In step 3, the specific process of constructing the classical domain matter-element matrix, the section domain matter-element matrix, the matter-element to be evaluated, and determining the real-time state numerical vector of the matter-element to be evaluated and the feature center vector of each risk level is as follows: Building a risk assessment set The safety risks of power transmission and transformation project construction are divided into: Each level is denoted as... ; Construct the first Classical domain matter-element matrix with risk levels :

[0017] In the formula, Indicates the first One risk level; Indicates the first One evaluation indicator; Indicators In the The range of values ​​under each risk level, among which and These are the lower and upper limits of the range, respectively. Constructing a domain matter-element matrix for risk assessment :

[0018] In the formula, Represents the entire risk domain; Indicators The range of allowed values ​​in the section domain, where and These are the lower and upper limits of the range, respectively. Constructing the object element to be evaluated in the sample to be evaluated :

[0019] In the formula, This indicates the specific engineering project to be evaluated; This indicates that the project meets the following criteria. The original measured data; Based on the segmented matter-element matrix The boundary values ​​for the original measured data Normalization process is performed to obtain : For positive indicators:

[0020] For negative indicators:

[0021] In the formula, Indicators In the domain element The lower and upper limits in the range; The normalized value represents the measured data of the object element to be evaluated; the normalized value ranges from [value missing]. The closer to 1, the safer it is; This allows us to obtain the real-time state numerical vector of the object to be evaluated. ; According to the classical field matter-element matrix The boundary value is calculated. Under the first risk level Item Indicators Interval characteristic median :

[0022] In the formula, and Indicators In the Classical Domain Matter Element under Each Risk Level The lower and upper limits in the value, This indicates that the original measured data was used. The same normalization rules are applied. Thus, the first Feature center vector of each risk level .

[0023] In step S4, the specific process of constructing the penalized state-weighted function and calculating the state-weighted vector and final weighted vector of the object to be evaluated is as follows: Construct the element to be evaluated Penalized state-variable weighting function for each index :

[0024] In the formula, The real-time state numerical vector of the object to be evaluated determined in step S3 The Middle The numerical components corresponding to each indicator; The preset penalty threshold is used to determine whether the indicator state has changed abruptly. The preset penalty intensity factor is used to adjust the degree of penalty imposed by the variable weight function on the deterioration index; Thus, the state-variable weight vector of the object element to be evaluated is obtained. ; Combining the aforementioned basic constant weight vector Calculate the element to be evaluated The final variable weight of each indicator :

[0025] In the formula, Basic constant weight vector The Middle The weight values ​​of each indicator; For the revised first Adaptive weighting of each indicator; This yields the final variable weight vector of the object element to be evaluated.

[0026] In step S5, the real-time state numerical vector of the object to be evaluated is calculated and compared with the first... The specific process of the weighted Mahalanobis distance between the feature center vectors of each risk level is as follows: Constructing the covariance matrix of evaluation indicators And calculate its inverse matrix. :

[0027] In the formula, For the first The variance of each indicator, For the first The first indicator and the first Covariance of each indicator; The inverse of the covariance matrix is ​​used to eliminate the interference of linear correlation between evaluation indicators; The final variable weight vector of the object to be evaluated obtained in step S4. Convert to a diagonal matrix :

[0028] Using the diagonal matrix and inverse matrix Calculate the real-time state numerical vector of the object to be evaluated and the first... The weighted Mahalanobis distance between the feature center vectors of each risk level :

[0029] In the formula, The real-time state numerical vector of the object to be evaluated, as determined in step S3. For the first step S3, the extracted Each risk level has a feature center vector.

[0030] In step S6, the specific process of calculating the closeness of the object to be evaluated to each risk level and determining the final risk level is as follows: Calculate the element to be evaluated with respect to the first The closeness of the risk levels :

[0031] In the formula, The real-time state numerical vector of the object to be evaluated calculated in step S5 is compared with the first... The variable-weight Mahalanobis distance between the feature center vectors of each risk level; The total number of risk levels; Based on the principle of maximum proximity, the risk level of the object to be evaluated is determined. :

[0032] If the object to be evaluated is related to the first The closeness of the risk levels If the risk level is the highest, then the final safety risk level of the power transmission and transformation project is determined to be [missing information]. .

[0033] Compared with the prior art, the beneficial effects of the present invention are: 1. Based on the framework of construction human factors analysis and classification system, a construction safety risk assessment index system for power transmission and transformation projects was constructed. Through data collection, expert demonstration, questionnaires, and data analysis, the system was screened from four dimensions: enterprise organizational influence, safety supervision, on-site operation-related factors, and construction personnel-related factors. Finally, a construction safety risk assessment index system for power transmission and transformation projects containing 15 key evaluation indicators was determined to further achieve a comprehensive and multi-faceted evaluation while maintaining a reasonable and scientific calculation workload. 2. Subjective and objective weights are calculated using the BWM method and the CRITIC method respectively, and then integrated through a game theory-based combined weighting method. This method retains the decision-making advantage of expert experience in dealing with qualitative indicators, while fully exploring the objective laws of the data itself, effectively avoiding the limitations of a single weighting method, and making the basic constant weight vector more scientific and reasonable.

[0034] 3. This invention introduces a punitive state-weighted mechanism. When a certain evaluation indicator deteriorates severely and exceeds a preset threshold, the model will automatically increase the weight of the indicator to amplify it punitively. This effectively overcomes the defect that key risk factors are easily neutralized by other good indicators under the traditional constant weight mode, and significantly improves the sensitivity of the evaluation method to identify local sudden safety hazards.

[0035] 4. The traditional matter-element extension model is improved by using variable weight Mahalanobis distance. By introducing the inverse covariance matrix to eliminate the interference of correlation between evaluation indicators, the true distance between the real-time state value vector of the matter to be evaluated and the feature center vector of each risk level is objectively restored, thereby significantly improving the accuracy and reliability of the safety risk level determination of power transmission and transformation engineering construction in complex environments. Attached Figure Description

[0036] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0037] Figure 1 A flowchart of a power transmission and transformation construction risk assessment method based on variable weight Mahalanobis distance matter-element extension provided for an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the structure of the safety risk assessment index system for power transmission and transformation engineering construction in an embodiment of the present invention. Detailed Implementation To make the objectives, technical solutions, and advantages of this invention more intuitive, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. The specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0038] Example 1 like Figure 1 As shown, this embodiment provides a method for risk assessment of power transmission and transformation construction based on variable-weight Mahalanobis distance matter-element extension. This method aims to solve the problems of evaluation distortion caused by the difficulty in eliminating interference from correlations between evaluation indicators and the inability of constant-weight evaluation models to sensitively reflect major local safety hazards in existing evaluation techniques. It eliminates interference from correlations by introducing the inverse covariance matrix of evaluation indicators and combines it with a penalized variable-weight mechanism for adaptive evaluation, thereby improving the accuracy and sensitivity of determining the safety risk level of power transmission and transformation construction. The method includes the following specific steps: Step S1: Introduce the Human Factors Analysis and Classification System (HFACS) framework to construct a safety risk assessment index system for power transmission and transformation projects. Verify the system through data collection, expert demonstration, questionnaires, and other methods, including positive, negative, and actual project feedback. Finally, select 15 evaluation indicators to form the final safety risk assessment index system for power transmission and transformation projects. In this embodiment, a safety risk assessment index system is first established based on the Construction Human Factors Analysis and Classification System (HFACS) framework, taking into account the characteristics of power transmission and transformation engineering construction operations. Through data collection, expert consultation, questionnaires, and data analysis, the system is screened from four dimensions: corporate organizational influence, safety supervision, on-site operation-related factors, and construction personnel-related factors. Ultimately, a safety risk assessment index system for power transmission and transformation engineering construction, comprising 15 key evaluation indicators, is determined. Figure 2 As shown.

[0039] The construction safety risk assessment index system for power transmission and transformation projects is divided into a two-tier structure: the first-tier index includes: the impact of enterprise organization. Safety supervision Factors related to on-site operations Factors related to construction personnel There are 15 secondary indicators in total, distributed as follows: (The remaining text appears to be a list of indicators related to organizational influence.) Organizational structure and responsibilities Investment in safe production Safety Management Procedures ; falls under the aforementioned safety supervision Risk monitoring and early warning Supervision and management violations Work plan arrangement Factors related to the aforementioned on-site operations Basic construction conditions Construction technical measures Construction machinery Construction natural environment Construction social environment Factors related to the construction personnel mentioned above illegal operations Skill errors Intuition and Decision Errors Personnel quality .

[0040] Step S2: Calculate the subjective and objective weights of each evaluation index using the BWM and CRITIC methods respectively, and then fuse the subjective and objective weights using a game-theoretic combination weighting method to obtain the basic constant weight vector of the evaluation index: In this embodiment, the subjective weights of the evaluation indicators are first calculated using the Brown-Wood Method (BWM). The specific process is as follows: (1) Determine the optimal and worst indicators: Based on the expert panel's opinion, the optimal indicator was determined from the evaluation indicator system established in step S1. and worst indicators Among them, the optimal indicator refers to the indicator that experts believe has the greatest impact on safety risks, while the worst indicator refers to the indicator that experts believe has the least impact on safety risks.

[0041] (2) Construct the optimal index comparison vector: Decision-makers will use the optimal indicators Pairwise comparisons are made with other indicators, using a 1-9 scale to determine the preference relationships, where a value of 1 represents that both are equally important, and a value of 9 represents that the optimal indicator is extremely important compared to the latter. This yields the optimal indicator comparison vector. : (1) In the formula, Indicates the optimal index relative to the first Importance preference values ​​for each indicator.

[0042] (3) Construct the worst-case index comparison vector: Similarly, decision-makers compare other indicators to the worst-case indicator. Perform pairwise comparisons to obtain the worst-case index comparison vector. : (2) In the formula, Indicates the first The importance preference value of each indicator relative to the worst indicator.

[0043] (4) Construct and solve the BM optimization model: In order to obtain the optimal subjective weight vector for each evaluation index To maximize the consistency of experts' logical judgments during pairwise comparisons, i.e., to minimize preference propagation error, the following minimization-maximization model is first constructed: (3) In the formula, Subjective weights representing the optimal indicator; Subjective weights representing the worst-performing indicator; Representing the Subjective weighting of each indicator.

[0044] To facilitate calculation, a consistency parameter is introduced. The above nonlinear model, i.e., formula (3), is transformed into the following linear BWM optimization model: (4) By solving the above linear BWM optimization model, i.e., formula (4), the optimal consistency parameters can be obtained. and the subjective weight vector of the evaluation indicators .

[0045] To verify the reliability of the expert scoring logic, a consistency ratio is introduced. To perform a consistency check: First, based on the maximum importance preference value used in the comparison vector, look up the corresponding consistency index in a table. Then, according to the formula Calculate the consistency ratio. If the consistency ratio is... If the value is less than the preset validity threshold, the consistency check is considered passed; otherwise, the comparison vector needs to be re-evaluated. and The model is revised to eliminate conflicts in the subjective logic of the experts until it passes the test, thereby ensuring the robustness of the evaluation model.

[0046] Furthermore, the CRITIC method is then used to calculate the objective weights of the evaluation indicators. The specific process is as follows: (1) Construct the initial indicator data matrix: Assume there is One sample to be evaluated. Evaluation indicators, and establish an initial indicator data matrix. .in, Represents the first in the initial evaluation matrix Line number The element of the column, i.e., the first The first of the samples to be evaluated Evaluation information for each indicator.

[0047] (2) Data standardization processing: To eliminate the influence of different indicator units on the results, the data is dimensionless to obtain a standardized matrix. .in, The normalized matrix obtained after normalizing the initial evaluation matrix is ​​the first normalized matrix. Line number Column elements, with a value range of 100. The closer the value is to 1, the safer it is.

[0048] For positive indicators, the normalization formula is as follows: (5) For negative indicators, the normalization formula is as follows: (6) (3) Calculate the standard deviation of each indicator: (7) In the formula, Indicates the first The standard deviation of the indicators; For the first The average value of each indicator.

[0049] (4) Calculate the conflict and information content of the indicators: First, calculate the Pearson correlation coefficient between the indicators. : (8) Based on the above Pearson correlation coefficient Calculate the first Conflicts among indicators : (9) Then calculate the first Information content of each indicator : (10) In the formula, As an indicator and indicators The Pearson correlation coefficient between them ranges from 1 to 10. Positive numbers represent a positive correlation between the two indicators, negative numbers represent a negative correlation between the two indicators, and 0 represents a linear and uncorrelated relationship between the two indicators.

[0050] (5) Determine the objective weight vector using the CRITIC model: (11) By solving the above formula (11), the objective weight vector of each evaluation index can be obtained. .

[0051] It is worth noting that the Pearson correlation coefficient used in the CRITIC method can only measure the linear coupling relationship between indicators. When the correlation coefficient is 0, it only indicates that there is no linear correlation between indicators, but complex nonlinear dependencies may exist. Given that the risks in power transmission and transformation projects often have the characteristics of nonlinear abrupt changes, relying solely on the linear correlation determination of the CRITIC method may miss some risk associations. Therefore, this invention, based on objective weighting, further introduces a punitive variable weighting mechanism to specifically capture and handle this nonlinear risk amplification effect, thus overcoming the limitations of traditional linear statistical methods.

[0052] Furthermore, in order to balance the logic of subjective judgment with the regularity of objective data and avoid the bias of a single weighting method, a game theory-based combined weighting method is finally used to fuse the above-mentioned subjective and objective weights to obtain the basic constant weight vector of the evaluation index. The specific process is as follows: (1) Construct a linear combination vector of subjective and objective weights: (12) In the formula, A linear combination vector representing subjective and objective weights; and These are the coefficients for the combination of subjective and objective weights, respectively. The optimal subjective weight vector of BWM obtained by solving equation (4); The objective weight vector of CRITIC is obtained by solving equation (11).

[0053] (2) Construct the objective function for minimizing the deviation: Based on the ideas of game theory models, in order to combine weights With each basic weight ( The total deviation between the two is minimized for the combination coefficients. , To optimize and find the optimal combination coefficients, the following optimization model is constructed: (13) In the formula, It represents the 2-norm of a vector.

[0054] (3) Solve for the optimal combination coefficients: Based on the properties of matrix differentiation, the above optimization problem is transformed into solving for the optimal combination coefficients that satisfy the first derivative condition. The system of linear differential equations: (14) (4) Coefficient normalization and determination of final weights: Solving the system of linear equations, i.e., formula (14), yields... and The coefficients are then normalized according to the following formula to obtain the final optimal linear combination coefficients. and Thus, the basic constant weight vector based on game theory weighting is calculated. : (15) After obtaining the basic constant weight vector, scientific methods are needed to quantify the degree of danger. Given that the safety evaluation of power transmission and transformation engineering construction is a complex system problem involving multiple parameters, multiple levels, and high uncertainty, the traditional matter-element extension model has significant limitations in such applications. On the one hand, the Euclidean distance used in traditional models implicitly assumes the independence of indicators, ignoring the strong correlation between evaluation indicators in actual working conditions, which easily leads to information redundancy and evaluation distortion. On the other hand, the fixed constant weight pattern cannot reflect the dynamic impact of indicator states, and is prone to numerical neutralization effects, thus masking local disaster-causing hazards.

[0055] To overcome the aforementioned shortcomings, this invention further constructs a matter-element extension evaluation model based on modified Mahalanobis distance. This model first introduces Mahalanobis distance to replace Euclidean distance, using the covariance matrix to eliminate correlation interference between evaluation indicators and objectively restore the true distance between the sample to be evaluated and the feature centers of each risk level. Secondly, it introduces a punitive state-modified weighting mechanism, dynamically adjusting the weights according to the real-time state of the indicators, and nonlinearly amplifying key indicators that touch the safety bottom line, effectively implementing the veto management concept in safety risk control. The evaluation model constructed in this invention effectively achieves a leap from linear static to nonlinear dynamic risk assessment, significantly improving the accuracy of the final evaluation results and the sensitivity to major local safety hazards.

[0056] Step S3: Construct the classical domain matter-element matrix and the section domain matter-element matrix for risk assessment, construct the matter-element to be assessed for the sample to be assessed, and normalize the measured data to be assessed based on the section domain matter-element boundary to determine the real-time state numerical vector of the matter-element to be assessed, and extract the feature center vector of each risk level from the classical domain matter-element matrix based on the same normalization rule. In this embodiment, the matter-element extension model for risk assessment is constructed, and the specific process is as follows: (1) Constructing a risk assessment set: Based on the construction characteristics of power transmission and transformation projects and relevant safety management regulations, construction safety risks are divided into 5 levels, denoted as set. .in represent Low risk level represent The risk level is relatively low. represent Level 1 Medium Risk represent High level of risk represent Level 1 High Risk.

[0057] (2) Constructing the classical domain matter-element matrix: The classical domain represents the range of values ​​for various indicators within a specific risk level. The first... Classical domain matter elements with risk levels as follows: (16) In the formula, Indicates the first Risk Level Classical domain elements; Indicates the first One risk level; Indicates the first One evaluation indicator; Indicates the first Under each risk level, the indicator The range of values ​​for .

[0058] (3) Construct the domain matter-element matrix: The node domain represents the total range of values ​​for all evaluation indicators within the entire evaluation system. Define the node domain matter element. as follows: (17) In the formula, A comprehensive object representing the construction safety risks of power transmission and transformation projects; Indicators The range of allowed values ​​throughout the evaluation, where and These are the lower and upper limits of the range, respectively.

[0059] (4) Construct the object element to be evaluated and determine the real-time state numerical vector: (18) In the formula, This indicates the specific engineering project to be evaluated; This indicates that the project meets the following criteria. The original measured data.

[0060] Based on the segmented matter-element matrix The boundary values ​​for the original measured data Normalization process is performed to obtain : For positive indicators: (19) For negative indicators: (20) In the formula, Indicators In the domain element The lower and upper limits in the range; The normalized value of the measured data in the object to be evaluated (the value range is 100%) (The closer to 1, the safer it is). By calculating the above formula, i.e., formula (19) or formula (20), the real-time state numerical vector of the object to be evaluated can be obtained. ; (5) Calculate the risk level feature center vector: According to the classical field matter-element matrix The boundary value is calculated. Under the first risk level Item Indicators Interval characteristic median : (twenty one) In the formula, and Indicators In the Classical Domain Matter Element under Each Risk Level The lower and upper limits in the value, This indicates that the original measured data was used. The same normalization rules are applied.

[0061] By calculating the above formula, i.e., formula (21), we can obtain the first... Feature center vector of each risk level .

[0062] Step S4: Construct a penalized state-weighted function. Based on the real-time state value vector of the object to be evaluated determined in Step S3, and combined with a preset penalty threshold and penalty intensity factor, construct a state-weighted vector of the object to be evaluated. Then, use the state-weighted vector to adaptively correct the basic constant weight vector obtained in Step S2, and calculate the final weighted vector of the object to be evaluated. In this embodiment, a penalized state-weighted function is constructed to obtain the state-weighted vector of the object to be evaluated, and the final weighted vector is calculated. The specific process is as follows: (1) Constructing a penalized state-change function: For each indicator in the indicator system, construct the first element of the object to be evaluated. Penalized state-variable weighting function for each index : (twenty two) In the formula, The real-time state numerical vector of the object to be evaluated determined in step S3 The Middle The numerical components corresponding to each indicator; The preset penalty threshold is used to determine whether the indicator state has changed abruptly. The preset penalty factor is used to adjust the degree of penalty imposed by the variable weight function on the deterioration index.

[0063] By calculating the above formula (22), the state-variable weight vector of the object to be evaluated is obtained. ; (2) Calculate the final variable weights: Combined with the basic constant weight vector obtained in step S2 Using the aforementioned state-change weight vector Treatment of the first element The weights of each indicator are adjusted, and the calculation formula is as follows: (twenty three) In the formula, For the revised first Adaptive weighting of each indicator; Basic constant weight vector The Middle The weight values ​​of each indicator.

[0064] By calculating the above formula (23), the final variable weight vector of the object to be evaluated can be obtained. .

[0065] This paper introduces a penalized state-weighted mechanism, enabling the model to adaptively amplify the weight of an evaluation indicator when its degradation exceeds a preset threshold. This mechanism effectively overcomes the shortcoming of the traditional constant-weight model, where key risks are easily neutralized by other good indicators, and significantly improves the sensitivity of the evaluation method in identifying localized sudden safety hazards.

[0066] Step S5: Construct the covariance matrix and its inverse matrix of the evaluation index, convert the final variable weight vector of the object to be evaluated into a diagonal matrix, and then use the covariance inverse matrix and the diagonal matrix to calculate the variable weight Mahalanobis distance between the real-time state value vector of the object to be evaluated and the feature center vector of each risk level. In this embodiment, the variable-weight Mahalanobis distance between the real-time state value vector of the object to be evaluated and the feature center vector of each risk level is calculated. The specific process is as follows: Constructing the covariance matrix of evaluation indicators And calculate its inverse matrix. : (twenty four) In the formula, For the first The variance of each indicator, For the first The first indicator and the first Covariance of each indicator; The inverse of the covariance matrix is ​​used to eliminate the interference of linear correlation between evaluation indicators.

[0067] (2) Construct a variable weight diagonal matrix: The final variable weight vector of the object to be evaluated obtained in step S4. Convert to a diagonal matrix : (25) (3) Calculate the variable-weight Mahalanobis distance: Using the diagonal matrix and inverse matrix Calculate the real-time state numerical vector of the object to be evaluated and the first... The weighted Mahalanobis distance between the feature center vectors of each risk level : (26) In the formula, The real-time state numerical vector of the object to be evaluated, as determined in step S3. For the first step S3, the extracted Each risk level has a feature center vector.

[0068] This paper improves the matter-element extension model by using variable-weight Mahalanobis distance instead of traditional Euclidean distance. By introducing the inverse covariance matrix into the distance measurement, the calculation process effectively eliminates the interference of inherent correlations among various evaluation indicators, objectively and accurately restoring the true spatial distance between the real-time state numerical vector of the matter to be evaluated and the feature center vector of each risk level. This significantly improves the accuracy and reliability of determining the safety risk level of power transmission and transformation engineering construction under complex working conditions.

[0069] Step S6: Calculate the proximity of the object to be evaluated with respect to each risk level based on the variable weight Mahalanobis distance, and determine the final risk level of the construction safety of the power transmission and transformation project based on the principle of maximum proximity.

[0070] In this embodiment, the degree of closeness of the object to be evaluated with respect to each risk level is calculated and the final risk level is determined. The specific process is as follows: (1) Calculate the closeness: Calculate the element to be evaluated with respect to the first The closeness of the risk levels : (27) In the formula, The real-time state numerical vector of the object to be evaluated calculated in step S5 is compared with the first... The variable-weight Mahalanobis distance between the feature center vectors of each risk level; The total number of risk levels; (2) Determine the final risk level: Based on the principle of maximum proximity, the risk level of the object to be evaluated is determined. : (28) If the object to be evaluated is related to the first The closeness of the risk levels If the risk level is the highest, then the final safety risk level of the power transmission and transformation project is determined to be [missing information]. .

[0071] Example 2 This embodiment selects a 220kV power transmission and transformation project in East China as the sample to be evaluated, and conducts a specific risk level assessment on it. The specific analysis steps are as follows: 1. Construct a safety risk assessment index system for power transmission and transformation engineering construction. In this embodiment, a safety risk assessment index system is first established based on the framework of construction human factors analysis and classification, taking into account the characteristics of power transmission and transformation engineering construction operations. Through data collection, expert consultation, questionnaires, and data analysis, the system is screened from four dimensions: enterprise organizational influence, safety supervision, on-site operation-related factors, and construction personnel-related factors. Finally, a power transmission and transformation engineering construction safety risk assessment index system containing 15 key evaluation indicators is determined. (See details...) Figure 2 .

[0072] 2. Calculation of basic constant weight vector (1) Calculation of subjective weights based on the BWM method Step 1: Determine the optimal and worst indicators.

[0073] Based on the established risk assessment index system, experts were organized to compare the importance of each evaluation index according to the actual project conditions and their professional knowledge. Based on relevant data and expert recommendations, the optimal index was determined from the 15 secondary indicators of the power transmission and transformation project construction safety risk assessment index system. For violations worst indicator For the social environment of construction .

[0074] Step 2: Construct the comparison vector.

[0075] Based on the 1-9 scale, the expert group provided the optimal index. Optimal indicator comparison vector relative to other indicators :

[0076] At the same time, other indicators are given relative to the worst indicator. worst-case index comparison vector :

[0077] Step 3: Solve for weights and consistency checks.

[0078] By solving the above BWM optimization model, i.e., formula (4), the subjective weight vector of the evaluation index is obtained. :

[0079]

[0080] Step 4: Consistency check.

[0081] The optimal consistency parameters obtained from the model solution Consulting the BWM standard consistency index table, we can see that when the maximum importance preference value is 9, the corresponding consistency index is... According to the formula The consistency ratio was calculated. .because This indicates that the comparison vector constructed by the expert group has extremely high consistency, and the weight calculation results are true and effective.

[0082] (2) Calculation of objective weights based on CRITIC In this embodiment, the acceptance evaluation data of 20 typical power transmission and transformation projects were selected as the historical sample set, and the standard deviation of each evaluation index was calculated according to formulas (5)-(10). Conflict and information content As shown in Table 1. Table 1: Calculation Results of Standard Deviation, Conflictality, and Information Content of Each Evaluation Indicator

[0083] By solving the CRITIC model (Equation 11) above, the objective weight vector of the evaluation index is obtained. :

[0084]

[0085] (3) Combinatorial weighting calculation based on game theory This embodiment utilizes a game theory-based combinatorial weighting method to fuse the aforementioned subjective and objective weights. Based on formulas (14)-(15), the final optimal linear combination coefficients are calculated. and The final fundamental constant weight vector :

[0086]

[0087] Table 2 Summary of Subjective and Objective Weights and Game Theory Combination Weights for Each Evaluation Indicator

[0088] 3. Construction of Matter-Element Extension Model and Extraction of Risk Level Feature Center Vector Constructing a classical domain matter-element matrix and the domain matter matrix The value ranges of each indicator in the classical domain matter-element matrix for the five risk levels and the value ranges of each indicator in the section domain matter-element matrix are shown in Table 3.

[0089] Table 3. Value Range Setting Table for Indicators in Classic Domains and Section Domains for Each Risk Level

[0090] Construct the object to be evaluated Among them, a 220kV power transmission and transformation project in East China was selected as the sample to be evaluated. The data of each indicator of the sample to be evaluated are shown in Table 4.

[0091] Table 4 Data of the Samples to be Evaluated

[0092] The data of the sample to be evaluated is normalized to determine the real-time state numerical vector of the object to be evaluated. :

[0093]

[0094] Using formula (21), the feature center vectors of each risk level are obtained. :

[0095]

[0096]

[0097]

[0098]

[0099]

[0100]

[0101]

[0102]

[0103]

[0104] 4. Variable weight calculation and adaptive weight adjustment Based on the characteristics of safety risks in power transmission and transformation engineering construction, penalty thresholds are set. Punishment intensity factor The state-variable weight vector of the object to be evaluated is obtained according to formula (22). :

[0105]

[0106] The fundamental constant weight vector is calculated using formula (23). The weights of the objects to be evaluated are adjusted to obtain the final variable weight vector of the objects to be evaluated. :

[0107]

[0108] 5. Calculation of variable weight Mahalanobis distance The final variable weight vector of the object to be evaluated is obtained. Convert to a diagonal matrix Combining the inverse covariance matrix The real-time state numerical vectors of the objects to be evaluated are calculated according to formula (26). With the 5 risk level feature center vectors The variable weight Mahalanobis distance between The calculation results are shown in Table 5: Table 5. Calculation Results of Variable Weight Mahalanobis Distance

[0109] 6. Final Risk Level Determination and Comparative Analysis According to the variable weight Mahalanobis distance Formula (27) is used to calculate the closeness of the object to be evaluated with respect to each risk level. The calculation results are shown in Table 6: Table 6 Proximity Calculation Results

[0110] Based on the principle of maximum proximity, the final risk level of the construction safety of this power transmission and transformation project is determined to be: Medium risk level.

[0111] This example sets up a control group for comparative analysis, using a basic constant weight vector without introducing a variable weighting mechanism. The Mahalanobis distance was calculated and the risk level was determined. The experimental group used the variable-weight Mahalanobis distance model described in this invention. The closeness calculation results and risk level determination of the two models are shown in Table 7.

[0112] Table 7 Comparison of Similarity and Risk Level between Control Group and Experimental Group

[0113] Comparative analysis results show that, under the control group (i.e., the constant weight model), the risk level of the sample to be evaluated is determined to be... The risk level is relatively low, the reason being that, although key indicators were manipulated illegally... While there is significant deterioration, the high scores of the remaining 14 indicators, under a fixed weighting system, dilute the impact of low scores in some indicators due to the neutralizing effect of numerous high-scoring indicators. This masks the significant safety hazards actually existing at the construction site and leads to distorted evaluation results. However, in the variable weighting model of this invention, the risk level of the sample to be evaluated is determined as follows: Level 1 medium risk, because the introduced punitive state-change mechanism is highly sensitive to violations. Safety Management Procedures and risk monitoring and early warning For the abnormal states of eight indicators, a penalty-type state-weighting function is used to adaptively amplify the weights of these indicators, with particular emphasis on the most severely degraded violations. The punitive increase in indicator weights ensured that the evaluation results were reasonably tilted towards the risk level of the deteriorating indicators, thus successfully correcting the misjudgment of the constant weight model. This comparative result fully verifies the sensitivity and reliability of the model constructed in this invention in handling local indicator deterioration, effectively avoiding missed risk assessments, and thus significantly improving the scientificity and accuracy of the safety risk assessment for power transmission and transformation engineering construction.

[0114] While the embodiments disclosed in this invention are as described above, the content is merely for the purpose of facilitating understanding of the invention and is not intended to limit the invention. Any person skilled in the art to which this invention pertains may make any modifications and variations in form and detail of the implementation without departing from the spirit and scope disclosed herein; however, the scope of patent protection for this invention shall still be determined by the scope defined in the appended claims.

Claims

1. A method for assessing the construction risk of power transmission and transformation based on variable-weight Mahalanobis distance matter-element extension, characterized in that, Includes the following steps: Step S1: Introduce a framework for analyzing and classifying human factors in construction to build a preliminary indicator system for evaluating the construction safety risks of power transmission and transformation projects. Through data collection, expert consultation, questionnaires, and data analysis, positive verification, reverse verification, and actual project feedback verification are conducted to select multiple evaluation indicators and form the final indicator system for evaluating the construction safety risks of power transmission and transformation projects. Step S2: Calculate the subjective and objective weights of each evaluation index using the BWM method and the CRITIC method respectively, and then fuse the subjective and objective weights using the game theory combination weighting method to obtain the basic constant weight vector of the evaluation index. Step S3: Construct the classical domain matter-element matrix, the section domain matter-element matrix, and the matter-element to be evaluated for the risk assessment sample. Normalize the measured data to be evaluated based on the section domain matter-element boundary to determine the real-time state value vector of the matter-element to be evaluated. Extract the feature center vector of each risk level from the classical domain matter-element matrix based on the same normalization rule. Step S4: Construct a penalized state-weighted function. Based on the real-time state value vector of the object to be evaluated, and combined with the preset penalty threshold and penalty intensity factor, construct the state-weighted vector of the object to be evaluated. Then, use the state-weighted vector to adaptively correct the basic constant weight vector and calculate the final weighted vector of the object to be evaluated. Step S5: Construct the covariance matrix and its inverse matrix of the evaluation index, convert the final variable weight vector of the object to be evaluated into a diagonal matrix, and then use the covariance inverse matrix and the diagonal matrix to calculate the variable weight Mahalanobis distance between the real-time state value vector of the object to be evaluated and the feature center vector of each risk level. Step S6: Based on the variable weight Mahalanobis distance, calculate the proximity of the object to be evaluated with respect to each risk level, and determine the final risk level of the construction safety of the power transmission and transformation project according to the principle of maximum proximity.

2. The method for risk assessment of power transmission and transformation construction based on variable-weight Mahalanobis distance matter-element extension as described in claim 1, characterized in that, In step S1, the indicators in the final power transmission and transformation project construction safety risk assessment index system are divided into two levels; The primary indicators include the impact of corporate organization, safety supervision, on-site operation-related factors, and construction personnel-related factors; The secondary indicators are as follows: organizational impact includes organizational structure and responsibilities, safety production investment, and safety management procedures; safety supervision includes risk monitoring and early warning, supervision and management violations, and work plan arrangements; on-site operation related factors include basic construction conditions, construction technology measures, construction machinery, construction natural environment, and construction social environment; and construction personnel related factors include violations of regulations, skill errors, intuition and decision-making errors, and personnel quality.

3. The method for risk assessment of power transmission and transformation construction based on variable-weight Mahalanobis distance matter-element extension as described in claim 1, characterized in that, In step S2, the specific process of calculating the subjective weights of the evaluation indicators using the BWM method is as follows: Determine the optimal indicator in the risk assessment indicator system and worst indicators ; Determine the importance preference of the optimal indicator to other indicators, and construct the optimal indicator comparison vector. ,in Indicates the optimal index relative to the first The importance preference of each indicator; Determine the importance preference of other indicators to the worst indicator, and construct a comparison vector for the worst indicator. ,in Indicates the first The importance preference of each indicator relative to the worst indicator; Construct the following linear BWM optimization model: In the formula, The subjective weight of the optimal indicator; The subjective weighting for the worst-case indicator; For the first Subjective weighting of each indicator; For consistency parameters; Therefore, the optimal consistency parameters can be obtained by solving the model. Subjective weights of evaluation indicators and subjective weight vector .

4. The method for risk assessment of power transmission and transformation construction based on variable-weight Mahalanobis distance matter-element extension as described in claim 3, characterized in that, To verify the reliability of the expert scoring logic, a consistency ratio is introduced. Perform a consistency check: Based on the maximum importance preference value used in the comparison vector, the corresponding consistency index is obtained by looking up the BWM Consistency Index (CI) mapping table. ; According to the formula Calculate the consistency ratio; if the consistency ratio... If the value is less than the preset validity threshold, the consistency check is considered passed; otherwise, the comparison vector needs to be re-evaluated. and The model is revised to eliminate conflicts in the subjective logic of the experts until it passes the test, thereby ensuring the robustness of the evaluation model.

5. The method for risk assessment of power transmission and transformation construction based on variable-weight Mahalanobis distance matter-element extension as described in claim 4, characterized in that, In step S2, the specific process of calculating the objective weights of the evaluation indicators using the CRITIC method is as follows: Build includes One sample Initial data matrix of each evaluation indicator The normalized matrix is ​​obtained by normalization. ; Calculate the first Standard deviation of each indicator and the The first indicator and the first Pearson correlation coefficient among the indicators ; Calculate the first according to the following formula. Conflicts among indicators and amount of information : Calculate the objective weights of each evaluation indicator. To obtain the objective weight vector .

6. The method for risk assessment of power transmission and transformation construction based on variable weight Mahalanobis distance matter-element extension as described in claim 5, characterized in that: In step 2, the specific process of calculating the basic constant weight vector using the game theory combinatorial weighting method is as follows: Construct a linear combination vector of subjective and objective weights ,in and These are the combination coefficients of subjective weights and objective weights, respectively; and These are the subjective weight vector and the objective weight vector, respectively. Based on the principle of minimizing deviation, a method for solving the optimal combination coefficients is constructed. Linear equation model: Solving the system of linear equations yields and The final combination coefficients are obtained by normalizing them according to the following formula. and : According to the formula The basic constant weight vectors of each evaluation index are calculated. .

7. The method for risk assessment of power transmission and transformation construction based on variable-weight Mahalanobis distance matter-element extension as described in claim 6, characterized in that: In step 3, the specific process of constructing the classical domain matter-element matrix, the section domain matter-element matrix, the matter-element to be evaluated, and determining the real-time state numerical vector of the matter-element to be evaluated and the feature center vector of each risk level is as follows: Building a risk assessment set The safety risks of power transmission and transformation project construction are divided into: Each level is denoted as... ; Construct the first Classical domain matter-element matrix with risk levels : In the formula, Indicates the first One risk level; Indicates the first One evaluation indicator; Indicators In the The range of values ​​under each risk level, among which and These are the lower and upper limits of the range, respectively. Constructing a domain matter-element matrix for risk assessment : In the formula, Represents the entire risk domain; Indicators The range of allowed values ​​in the section domain, where and These are the lower and upper limits of the range, respectively. Constructing the object element to be evaluated in the sample to be evaluated : In the formula, This indicates the specific engineering project to be evaluated; This indicates that the project meets the following criteria. The original measured data; Based on the segmented matter-element matrix The boundary values ​​for the original measured data Normalization process is performed to obtain : For positive indicators: For negative indicators: In the formula, Indicators In the domain element The lower and upper limits in the range; The normalized value represents the measured data of the object element to be evaluated; the normalized value ranges from [value missing]. The closer to 1, the safer it is; This allows us to obtain the real-time state numerical vector of the object to be evaluated. ; According to the classical field matter-element matrix The boundary value is calculated. Under the first risk level Item Indicators Interval characteristic median : In the formula, and Indicators In the Classical Domain Matter Element under Each Risk Level The lower and upper limits in the value, This indicates that the original measured data was used. The same normalization rules are applied. Thus, the first Feature center vector of each risk level .

8. The method for risk assessment of power transmission and transformation construction based on variable weight Mahalanobis distance matter-element extension as described in claim 7, characterized in that, In step S4, the specific process of constructing the penalized state-weighted function and calculating the state-weighted vector and final weighted vector of the object to be evaluated is as follows: Construct the element to be evaluated Penalized state-variable weighting function for each index : In the formula, The real-time state numerical vector of the object to be evaluated determined in step S3 The Middle The numerical components corresponding to each indicator; The preset penalty threshold is used to determine whether the indicator state has changed abruptly. The preset penalty intensity factor is used to adjust the degree of penalty imposed by the variable weighting function on the deterioration index; Thus, the state-variable weight vector of the object element to be evaluated is obtained. ; Combining the aforementioned basic constant weight vector Calculate the element to be evaluated The final variable weight of each indicator : In the formula, Basic constant weight vector The Middle The weight values ​​of each indicator; For the revised first Adaptive weights for each indicator; This yields the final variable weight vector of the object element to be evaluated. .

9. The method for risk assessment of power transmission and transformation construction based on variable-weight Mahalanobis distance matter-element extension as described in claim 8, characterized in that, In step S5, the real-time state numerical vector of the object to be evaluated is calculated and compared with the first... The specific process of the weighted Mahalanobis distance between the feature center vectors of each risk level is as follows: Constructing the covariance matrix of evaluation indicators And calculate its inverse matrix. : In the formula, For the first The variance of each indicator, For the first The first indicator and the first Covariance of each indicator; The inverse of the covariance matrix is ​​used to eliminate the interference of linear correlation between evaluation indicators; The final variable weight vector of the object to be evaluated obtained in step S4. Convert to a diagonal matrix : Using the diagonal matrix and inverse matrix Calculate the real-time state numerical vector of the object to be evaluated and the first... The weighted Mahalanobis distance between the feature center vectors of each risk level : In the formula, The real-time state numerical vector of the object to be evaluated, as determined in step S3. For the first step S3, the extracted Each risk level has a feature center vector.

10. The method for risk assessment of power transmission and transformation construction based on variable-weight Mahalanobis distance matter-element extension as described in claim 9, characterized in that, In step S6, the specific process of calculating the closeness of the object to be evaluated to each risk level and determining the final risk level is as follows: Calculate the element to be evaluated with respect to the first The closeness of the risk levels : In the formula, The real-time state numerical vector of the object to be evaluated calculated in step S5 is compared with the first... The variable-weight Mahalanobis distance between the feature center vectors of each risk level; The total number of risk levels; Based on the principle of maximum proximity, the risk level of the object to be evaluated is determined. : If the object to be evaluated is related to the first The closeness of the risk levels If the risk level is the highest, then the final safety risk level of the power transmission and transformation project is determined to be [missing information]. .

Images