Power distribution automation terminal state comprehensive evaluation method based on AKFCM model

Abstract

The invention relates to a power distribution automation terminal state comprehensive evaluation method based on an AKFCM model, and the method comprises the steps: constructing a power distribution automation terminal state evaluation index system, combining the subjective weight and objective weight of a comprehensive evaluation index, calculating a combined weight through employing a game theory, and determining an adaptive kernel of an evaluation grade. Comprehensively evaluating the state of the distribution automation terminal based on an adaptive kernel fuzzy C-means clustering AKFCM model, calculating an index value of a current clustering result by adopting a multi-attribute decision-making method VIKOR method, comparing the index value with an ideal state, and performing evaluation to obtain a final evaluation grade; the subjective weight and the objective weight of the comprehensive evaluation index are combined, the game theory is utilized to calculate the combined weight, the limitation of a single weighting method is overcome, the self-adaptive kernel fuzzy C-means clustering model is introduced to process the fuzziness and the uncertainty of the power distribution network index, the clustering effect is improved, the closeness of the clustering result and the ideal state is evaluated, and the power distribution network performance is improved. The state level of the power distribution automation terminal can be evaluated more accurately.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution networks, and particularly to a comprehensive evaluation method for the status of distribution automation terminals based on the AKFCM model. Background Art

[0002] At present, the evaluation research related to distribution networks mostly focuses on aspects such as investment, construction, and risk assessment. There is less evaluation research on its dispatching operation management level. What can be referred to is distribution network planning. The method of relying on expert experience to evaluate the power supply mode of important power users in the distribution network is too dependent on expert subjective opinions and is prone to artificial subjective extreme value deviations. There are many objective weight determination methods, which overcome the subjectivity of expert assignment to a certain extent. However, the objective weight assignment is too dependent on measured index data and does not consider the actual importance of the indexes. Moreover, in the dispatching operation management of the distribution network, many indexes have fuzziness and uncertainty. For example, some indexes may be difficult to accurately quantify, or there may be differences in the evaluation of certain indexes by different experts.

[0003] The evaluation of the status of distribution automation terminal equipment is a practical problem that urgently needs to be solved on site. However, due to the uncertainty of some of the evaluation indexes and the fuzziness, randomness, and grayness of the evaluation grade standards, conventional evaluation methods are difficult to obtain acceptable evaluation conclusions, so it has not been well solved. Therefore, this patent proposes a comprehensive evaluation method for the status of distribution automation terminal equipment. Summary of the Invention

[0004] The present invention provides a comprehensive evaluation method for the status of distribution automation terminals based on the AKFCM model. By combining the subjective weight and objective weight of the comprehensive evaluation indexes, using game theory to calculate the combined weight, overcoming the limitations of the single weighting method, introducing an adaptive kernel fuzzy C-means clustering model to process the fuzziness and uncertainty of distribution network indexes, improving the clustering effect, and evaluating the closeness of the clustering result to the ideal state, more accurate evaluation of the status level of distribution automation terminals is realized.

[0005] To achieve the above object, the present invention is implemented by adopting the following technical solutions:

[0006] A comprehensive evaluation method for the status of distribution automation terminals based on the AKFCM model includes the following steps:

[0007] S1. Construct a status evaluation index system for distribution automation terminals;

[0008] S2. Calculate the comprehensive evaluation index weights for the distribution automation terminal status: Construct a judgment matrix and determine the subjective weights of the comprehensive evaluation indicators through the analytic hierarchy process. At the same time, standardize the data indicators and perform consistency verification. Then, use the ordered weighted averaging (OWA) operator to correct the subjective weights and the entropy weight method to determine the objective weights of the comprehensive evaluation indicators, and determine the final combined weights through game theory combined weighting;

[0009] S3. Determine the adaptive kernel of the evaluation level, comprehensively evaluate the distribution automation terminal status based on the adaptive kernel fuzzy C-means clustering (AKFCM) model, use the multi-attribute decision-making method VIKOR method to calculate the index values of the current clustering results and compare them with the ideal state for evaluation, and obtain the final evaluation level.

[0010] Furthermore, the distribution automation terminal status evaluation index system includes first-level indicators and second-level indicators, and the first-level indicators include four second-level indicators;

[0011] The first-level indicators include equipment performance, communication quality, control function, and reliability;

[0012] The first-level indicator of equipment performance includes second-level indicators of appearance integrity, installation stability, electrical parameter accuracy, and insulation performance level;

[0013] The first-level indicator of communication quality includes second-level indicators of communication stability, data transmission rate, and signal strength;

[0014] The first-level indicator of control function includes second-level indicators of remote control response time, remote control execution accuracy rate, and local control effectiveness;

[0015] The first-level indicator of reliability includes second-level indicators of mean time between failures, fault repair timeliness, equipment aging degree assessment, and availability of spare parts.

[0016] Furthermore, the steps for determining the subjective weights of the comprehensive evaluation indicators are as follows:

[0017] S2.11. According to expert opinions, through the analytic hierarchy process and after consistency verification, obtain the initial weight matrix B(n×s) of s experts for n evaluation indicators. For the weights (b j1 , b j2 , b j3 , …, b js ) obtained for the jth indicator, perform a descending order rearrangement, denoted as (p 0 , p 1 , …, p s-1 ). Select a data combination after the rearrangement as Then the formula for the ordered weighted vector vt is as follows:

[0018]

[0019] Among them, t is the sequence number of the data in the combination after selecting one data from the rearranged data; in the

[0020] S2.12. The absolute weight w of the j-th index j is calculated as follows:

[0021]

[0022] Among them, p ι is the evaluation of the i-th expert for the j-th index;

[0023] S2.13. Use the ordered weighted operator OWA to correct the subjective weight:

[0024]

[0025] Furthermore, the steps for the entropy weight method to determine the objective weight of the comprehensive evaluation index are as follows:

[0026] S2.21. Use the cloud model to establish the cloud comments of multiple experts. Each language description corresponds to a cloud model, and the qualitative index comprehensive cloud comments are obtained by synthesizing the comments of multiple experts. The expectation in the qualitative index cloud comments is used as the final quantitative value of the expert group for the qualitative index;

[0027] S2.22. Calculate the objective weight of the evaluation index:

[0028]

[0029] Among them, h j is the objective weight vector of the evaluation index, and e j is the entropy value of the j-th index.

[0030] Furthermore, the final combined weight is determined by game theory combined weighting:

[0031]

[0032] Among them, ω is the combined weight vector, w j is the weight vector of the subjective weight, h j is the weight vector of the objective weight, α 1 is the combined coefficient of the subjective weight, α 2 is the combined coefficient of the objective weight, β 1 is the combined coefficient of the subjective weight after the game, β 2 is the combined coefficient of the objective weight after the game.

[0033] Further, the multi-attribute decision-making method VIKOR method is used to calculate the index value of the current clustering result and compare it with the ideal state for evaluation, including the following steps:

[0034] S2.31. Establish an initial evaluation matrix A: A = [a pq n×m, where q = (1, 2,..., n), p = (1, 2,..., m), m is the number of evaluation objects, n is the number of evaluation indicators, and a pq is the pth evaluation indicator value of the qth object to be evaluated;

[0035] S2.32. Establish a normalized matrix B and perform dimensionless processing on the initial evaluation matrix A:

[0036] For the benefit-type indicators where the larger the value, the better,

[0037]

[0038] For the cost-type indicators where the smaller the value, the better,

[0039]

[0040] where b pq is the dimensionless form of a pq , are respectively the minimum and maximum values of the elements in the qth column of the initial evaluation matrix A;

[0041] Determine the positive ideal solution and the negative ideal solution

[0042]

[0043] Statistically calculate the team performance value S p and the individual regret value R of all objects to be evaluated:

[0044]

[0045]

[0046] Calculate the benefit value Q p of each object to be evaluated:

[0047]

[0048] where λ is the decision-making mechanism coefficient, λ ∈ [0, 1], λ > 0.5 indicates decision-making based on maximizing the group benefit; λ < 0.5 indicates decision-making based on minimizing individual regret.

[0049] Compared with the prior art, the beneficial effects of the present invention are:

[0050] 1) The comprehensive evaluation of comprehensive evaluation indicators is carried out by a method combining subjective and objective. At the same time, the combination of subjective and objective weighting methods can better explore the relative relationships between data, which can not only reflect the professional evaluation of the status of distribution automation terminals, but also reveal the objective reality of actual operation data, so as to obtain more objective and accurate weights and overcome the limitations of single weighting methods.

[0051] 2) Game theory is used to calculate the combined weights of indicators, and the VIKOR method is applied to evaluate the closeness of these ratings to the ideal state, so that the evaluation results are not only professionally authoritative but also can accurately reflect the actual operation status of the terminals, providing a more scientific and reasonable basis for the maintenance, upgrade and transformation of equipment.

[0052] 3) An adaptive kernel fuzzy C-means clustering model is introduced to handle the fuzziness and uncertainty of distribution network indicators, and the closeness of the clustering results to the ideal state is evaluated. By introducing the concept of membership degree, allowing a data point to belong to multiple clusters with a certain membership degree, it can better handle this fuzziness and uncertainty, improve the clustering effect, and achieve a more accurate evaluation of the status level of distribution automation terminals. Different from traditional hard clustering methods, it does not strictly divide each data point into one category, but is more in line with the complex characteristics of data in actual situations. Brief Description of the Drawings

[0053] Figure 1 is the flowchart of the method of the present invention.

[0054] Figure 2 is the schematic diagram of the structure of the comprehensive evaluation index system for the status of distribution automation terminals of the present invention. Detailed Embodiments

[0055] The following further describes the detailed embodiments of the present invention with reference to the drawings:

[0056] See Figure 1 , which is the flowchart of the method of the present invention. A comprehensive evaluation method for the status of distribution automation terminals based on the AKFCM model of the present invention first conducts actual measurement statistics and calculations on quantitative indicators, and verbalizes qualitative indicators. Then, a judgment matrix is constructed and the subjective weights are determined by the analytic hierarchy process. At the same time, the data indicators are standardized and consistency verified. Then, the ordered weighted averaging (OWA) operator is used to modify the subjective weights, the entropy weight method is used to determine the objective weights, and the final weights are determined by game theory combined weighting. Next, the adaptive kernel of the evaluation level is determined, and the final weights are obtained by using the fuzzy C-means clustering model and the comprehensive evaluation clustering coefficient based on the adaptive kernel fuzzy C-means. Finally, the VIKOR method is used to compare with the ideal state for evaluation to obtain the final evaluation level, which specifically includes the following steps:

[0057] S1. Construct an evaluation index system for the status of distribution automation terminals;

[0058] Based on the actual operation of distribution automation terminals, construct a comprehensive evaluation index system for the status of distribution automation terminals according to the principles of comprehensiveness, scientificity, independence, etc. This system comprehensively covers all aspects of equipment performance, forming a complete system from the target layer to the criterion layer and then to the index layer to comprehensively evaluate the terminal status in all directions. The evaluation indicators are based on objective facts and data, and the measurement methods and standards are clarified to ensure the accuracy and reliability of the results. The indicators are independent of each other, avoiding duplication and intersection, and each indicator has a clear purpose and function.

[0059] Based on the above construction principles, through hierarchical analysis, divide the comprehensive evaluation indicators of the status of distribution automation terminals into three layers and establish a comprehensive evaluation index system; see Figure 2 , the first layer is the target layer, which is used to evaluate the terminal status level; the second layer is the criterion layer, including 4 first-level indicators of equipment performance, communication quality, control function, and reliability; the third layer is the index layer, and 14 second-level indicators are determined according to the criterion layer. The 14 second-level indicators are as follows:

[0060] 1) Appearance integrity: Evaluate whether the shell of the terminal device is damaged, cracked, deformed, etc., and check whether components such as the display screen and indicator lights are intact to ensure that the operating status and alarm information of the device can be normally displayed;

[0061] 2) Installation stability: Examine whether the device is firmly installed and whether the installation bolts are tightened to avoid the device loosening or displacing due to factors such as vibration and external forces during operation;

[0062] 3) Electrical parameter accuracy: Analyze the measurement accuracy of electrical parameters such as voltage, current, and power to ensure that the measured values of the device are within the allowable accuracy range;

[0063] 4) Insulation performance level: Measure the insulation resistance value and insulation level of the device to ensure that electrical faults such as leakage and short circuits do not occur during operation;

[0064] 5) Communication stability: Measure the communication interruption frequency, anti-interference ability, etc.;

[0065] 6) Data transmission rate: The speed at which data is transmitted between the terminal and the master station system. A higher data transmission rate can enable the system to respond to changes faster and improve real-time performance;

[0066] 7) Signal strength: A high signal strength means good communication quality and more reliable data transmission;

[0067] 8) Remote control response time: The time elapsed from when the master station system issues a remote control command to when the distribution automation terminal actually executes the command. A short response time can improve the control efficiency of the system and quickly respond to faults and changes;

[0068] 9) Remote control execution accuracy rate: Measures the proportion of remote control instructions that are correctly executed. A high execution accuracy rate ensures the reliable and effective control operations of the system and avoids misoperations.

[0069] 10) Local control effectiveness: Evaluates the ability and effect of the terminal to perform control operations locally.

[0070] 11) Mean time between failures (MTBF): Reflects the quality and design level of the equipment. A long MTBF indicates high equipment stability and less likelihood of failures.

[0071] 12) Fault repair timeliness: The time required from the occurrence of a fault in the equipment to the completion of the repair. Timely fault repair can reduce the power outage time and improve the availability of the system.

[0072] 13) Equipment aging degree assessment: Evaluates the aging state of the equipment and predicts the remaining service life of the equipment. Timely detection of severely aging equipment and replacement can avoid system failures caused by equipment faults.

[0073] 14) Spare parts availability: Includes aspects such as the types, quantities, and storage management of spare parts to ensure that the required spare parts can be provided in a timely manner for maintenance when equipment fails.

[0074] S2. Comprehensive evaluation index weight calculation for the distribution automation terminal status

[0075] Determining the index weights is the key to comprehensive evaluation. First, use the Analytic Hierarchy Process (AHP) to subjectively assign weights to relevant indicators, and then use the Ordered Weighted Averaging (OWA) operator to correct the subjective weights to reduce the impact of human subjective extreme deviations in the AHP on the weight accuracy. The main limitation of the AHP is that it is difficult to maintain objectivity, especially when the number of indicators at the evaluation level is large, such as more than four. Therefore, the entropy weight method is used to allocate index weights, which can enhance the objectivity of the evaluation. The entropy weight method determines weights based on the amount of information provided by different indicators. At the same time, combining subjective and objective weighting methods can better explore the relative relationships between data, thus obtaining more objective and accurate weights. Use game theory to calculate the combined weights of indicators.

[0076] S2.1. Determine the subjective weights

[0077] Taking the criterion layer indicators as an example, according to expert opinions, through the Analytic Hierarchy Process and after consistency testing, obtain the initial weight matrix B(n×s) of s experts for n evaluation indicators. For the weights obtained for the jth indicator (b j1 ,b j2 ,b j3 ,…,b js ), perform a descending order rearrangement, denoted as (p 0 ,p1 ,…, p s-1 ), the combination after selecting one data from the rearranged data is Then the formula for the ordered weighted vector vt is as follows:

[0078]

[0079] where t is the sequence number of this data in the combination after selecting one data from the rearranged data in

[0080] Thus, the absolute weight w j value of the j-th index is:

[0081]

[0082] where p ι is the evaluation of the i-th expert on the j-th index;

[0083] Then the subjective weight value w j after the OWA operator is corrected is:

[0084]

[0085] S2.2. Determining the objective weight by the entropy weight method

[0086] S2.21. Standardizing the evaluation index data

[0087] The entropy weight method is an objective weighting method that determines the index weight according to the degree of chaos of the index, that is, the amount of information; before calculating the objective weight, it is necessary to standardize the evaluation index data of the distribution network dispatching operation management level to ensure the same dimension. There are quantitative and qualitative indexes in the index system. For qualitative indexes, expert scoring is often used for quantification. Because of the lack of information, experts often give qualitative natural language descriptions and it is difficult to give specific scores. Considering that the cloud model can realize the uncertain conversion between qualitative language concepts and quantitative values and can integrate and reflect the fuzziness of the evaluation grade information and the randomness of the evaluator's subjective judgment, the present invention uses the cloud model to establish the cloud comments of multiple experts. Each language description corresponds to a cloud model, and the comprehensive cloud comments of qualitative indexes are obtained by synthesizing the comments of multiple experts. Among them, the expectation in the cloud comments of qualitative indexes is used as the final quantitative value of the expert group for qualitative indexes. For quantitative indexes, based on the literature standardization method, different types of indexes are standardized, and finally a standardized decision matrix of evaluation index data with multiple county dispatching and multiple evaluation indexes (Ex p , En p , Ee p ) is obtained.

[0088]

[0089] Among them, Ex p , En p , Ee p are variables related to different types of evaluation indicators, representing the expected value, value, and entropy value under certain specific conditions respectively. Ex, Ex f , En f are respectively, Ex is the expected value of a certain indicator, Ex f is the expected value related to a certain final state, and En f represents the final entropy value.

[0090] For quantitative indicators, based on the standardization processing method, the extremely large type indicators, interval type indicators, and extremely small type indicators are standardized. Finally, the standardized decision matrix X of the evaluation indicator data with m county-level dispatching and n evaluation indicators is obtained as follows:

[0091] X = (x ij ) m×n

[0092] Among them, x ij is the jth indicator of the ith county-level dispatching.

[0093] S2.22. Calculating the objective weight of indicators by the entropy weight method

[0094] The proportion b ij occupied by the jth indicator value x ij under the ith county-level dispatching is:

[0095]

[0096] Then the entropy value e j of the jth indicator is:

[0097]

[0098] Calculate the entropy weight of the jth indicator, that is, the objective weight vector h j is:

[0099]

[0100] S2.23. The combined weight based on game theory

[0101] By combining subjective and objective weighting methods, it can not only reflect the professional evaluation of the status of distribution automation terminals but also reveal the objective reality of actual operation data. To balance the subjective and objective weight coefficients, game theory is used to establish the weights of various indicators in the distribution automation terminal status evaluation system. This can make the evaluation results not only have professional authority but also accurately reflect the actual operation status of the terminals, providing a more scientific and reasonable basis for equipment maintenance, upgrading, and transformation. The core of this method lies in taking into account both subjective and objective factors, striving for consistency or compromise between the two, so as to minimize the weight difference. The combined weight formula for each indicator is as follows:

[0102] ω = α 1 w j + α 2 h j (19)

[0103] Among them, ω is the combined weight vector, α 1 is the combined coefficient of the subjective weight, and α 2 is the combined coefficient of the objective weight;

[0104] According to the principle of game theory, to minimize the deviation between the subjective and objective weights and the combined coefficients, that is

[0105] min||ω = α 1 w j + α 2 h j || (20)

[0106] According to matrix differential theory, derivative calculations are carried out to find that the condition for the optimal first derivative is:

[0107]

[0108] Solve the system of equations to obtain the combined coefficients of the subjective and objective weights, and then calculate the value of the subjective and objective combined weight vector after normalization:

[0109]

[0110] Among them, w j is the weight vector of the subjective weight, h j is the weight vector of the objective weight, β 1 is the combined coefficient of the subjective weight after the game, and β 2 is the combined coefficient of the objective weight after the game.

[0111] S3. Comprehensive evaluation of the status of distribution automation terminals based on the adaptive kernel fuzzy C - means clustering AKFCM model;

[0112] This model is an improvement on the traditional fuzzy C-means clustering. It introduces an adaptive kernel function that can automatically adjust the shape and structure of the clusters according to the data distribution. When calculating the membership degree and cluster centers, it takes into account the complex relationships between data points. By mapping the data to a high-dimensional space through the kernel function, it becomes easier to find a suitable cluster structure in the high-dimensional space. The steps are as follows:

[0113] S3.1. Data Preparation

[0114] There is a dataset X = {x 1 , x 2 ,..., x n} containing n data points related to the status of distribution automation terminals. Each data point contains multiple attributes in the comprehensive evaluation index system of distribution automation terminals. From the perspective of the 14 secondary indicators at the index layer, each data point can be represented as X i = {x i1 , x i2 ,..., x i14}, where each attribute corresponds to the value of a secondary indicator. Through the analytic hierarchy process, entropy weight method, and combination weight calculation based on game theory, the combination weight vector w = (w 1 , w 2 , …, w 14 ) of each secondary indicator is obtained.

[0115] S3.2. Initialization of Cluster Centers and Membership Matrix

[0116] Randomly select C data points as the initial cluster centers, denoted as V = {v 1 , v 2 , …, v C}, where v j = (v j1 , v j2 , …, v j14 ), j = 1, 2,..., C. Here, the attributes of each cluster center also correspond to the values of 14 secondary indicators. Initialize the membership matrix as:

[0117]

[0118] where u ij represents the degree to which the data point belongs to the cluster, V is the set of cluster centers, and v j is the cluster center, and it needs to satisfy:

[0119]

[0120] S3.3. Iteration Process

[0121] For each data point x iand the clustering center v j , calculate the weighted distance d ij , according to the weight vector w obtained by the subjective and objective weighting method, calculate the weighted distance using the following formula:

[0122]

[0123] The attributes here are the secondary indicators based on the status level of the distribution automation terminal, so the summation is from k = 1 to 14;

[0124] Select an adaptive kernel function such as the Gaussian kernel function, and its form is:

[0125]

[0126] where, K(x i , x j ) is the adaptive kernel function, and σ is the bandwidth parameter of the kernel function;

[0127] It is the bandwidth parameter of the kernel function:

[0128]

[0129] This can make the secondary indicators with higher weights have a greater impact on the kernel function, so as to better capture the clustering structure of the distribution automation terminal status data, and then perform membership degree and clustering update:

[0130]

[0131]

[0132] S3.4. Evaluation of clustering results

[0133] Use the VIKOR method to evaluate the closeness of the clustering result to the ideal state; first, it is necessary to determine the relevant index values of the distribution automation terminal status level indicators in the ideal clustering state, and the ideal clustering center position can be set, that is, the values and membership degree distributions of each secondary indicator in the ideal clustering state;

[0134] Evaluation with the ideal state based on the application of the multi-attribute decision-making method VIKOR method:

[0135] Mainly evaluate the proximity level of these alternatives to the ideal alternative by measuring the collective utility, regret degree and benefit degree of each alternative, so as to judge the advantages and disadvantages of the alternatives, including the following steps:

[0136] (1) Establish the initial evaluation matrix A: Assume there are m evaluation objects and n evaluation indicators, then the n×m-order initial evaluation matrix A = [a pqn×m, where q = (1, 2,..., n), p = (1, 2,..., m), and a pq is the p-th evaluation index value of the q-th object to be evaluated;

[0137] (2) Establish a normalized matrix B to perform dimensionless processing on the initial evaluation matrix A;

[0138] For benefit-type indicators where the larger the value, the better:

[0139]

[0140] For cost-type indicators where the smaller the value, the better:

[0141]

[0142] where b pq is the dimensionless form of a pq , and are respectively the minimum and maximum values of the elements in the q-th column of the initial evaluation matrix A.

[0143] (3) Determine the positive ideal solution and the negative ideal solution

[0144]

[0145] (4) Statistically calculate the team performance value S p and the individual regret value R of all objects to be evaluated:

[0146]

[0147]

[0148] (5) Calculate the benefit value Q p of each object to be evaluated:

[0149]

[0150] where λ is the decision-making mechanism coefficient, λ ∈ [0, 1]. When λ > 0.5, it means making a decision based on maximizing the group benefit; when λ < 0.5, it means making a decision based on minimizing the individual regret. Take λ = 0.5.

[0151] For the current clustering result, calculate the corresponding index values, compare them with the ideal state, and calculate indicators such as the collective utility value, regret degree value, and benefit value according to the calculation rules of the VIKOR method to evaluate the quality of the clustering result.

[0152] The above embodiments are implemented on the premise of the technical solution of the present invention, and detailed implementation manners and specific operation processes are given. However, the protection scope of the present invention is not limited to the above embodiments. The methods used in the above embodiments are all conventional methods unless otherwise specified.

Claims

1. A comprehensive evaluation method for distribution automation terminal status based on AKFCM model, characterized in that: The steps include: S1. Construct a distribution automation terminal status evaluation index system; S2. Calculate the weights of comprehensive evaluation indicators of distribution automation terminal status: construct a judgment matrix and determine the subjective weights of comprehensive evaluation indicators through the hierarchical analysis method, standardize and check the consistency of data indicators, and then use the ordered weighted operator OWA operator to correct the subjective weights and the entropy weight method to determine the objective weights of comprehensive evaluation indicators, and determine the final combined weights through game theory combination weighting; S3. Determine the adaptive kernel of the evaluation level, comprehensively evaluate the status of the distribution automation terminal based on the adaptive kernel fuzzy C-means clustering AKFCM model, and use the multi-attribute decision-making method VIKOR method to calculate the index value of the current clustering result and compare it with the ideal state to obtain the final evaluation level.

2. According to claim 1, a comprehensive evaluation method for distribution automation terminal status based on AKFCM model is characterized in that: The distribution automation terminal status evaluation index system includes primary indicators and secondary indicators, and the primary indicators include four secondary indicators; The first-level indicators include equipment performance, communication quality, control function and reliability; The first-level indicator equipment performance includes the second-level indicators of appearance integrity, installation stability, electrical parameter accuracy and insulation performance level; The first-level indicator communication quality includes the second-level indicators communication stability, data transmission rate and signal strength; The first-level indicator control function includes the second-level indicators remote control response time, remote control execution accuracy and local control effectiveness; The first-level indicator reliability includes the second-level indicators of mean time between failures, timeliness of fault repair, equipment aging assessment and availability of spare parts.

3. According to claim 1, a comprehensive evaluation method for distribution automation terminal status based on AKFCM model is characterized in that: The steps of determining the subjective weight of the comprehensive evaluation index include the following: S2.

11. Based on the expert opinions, the initial weight matrix B(n×s) of n evaluation indicators by s experts is obtained through the hierarchical analysis method and consistency test. The weight (b j1 ,b j2 ,b j3 ,…,b js ) are rearranged in descending order and recorded as (p0, p1, ..., p s-1 ), the combination after selecting one data in the rearranged data is Then the formula of ordered weighted vector vt is as follows: Among them, t is the combination after selecting a data from the rearranged data. The sequence number of the data in S2.12, the absolute weight w of the jth indicator j calculate: Among them, p ι is the evaluation of the i-th expert on the j-th indicator; S2.13, use the ordered weighted operator OWA to correct the subjective weight:

4. According to claim 1, a comprehensive evaluation method for distribution automation terminal status based on AKFCM model is characterized in that: The entropy weight method for determining the objective weight of the comprehensive evaluation index includes the following steps: S2.

21. Use cloud models to establish cloud comments from multiple experts. Each language description corresponds to a cloud model. The comments from multiple experts are combined to obtain a comprehensive cloud comment on the qualitative indicators. The expectations in the cloud comments on the qualitative indicators serve as the final quantitative values ​​of the qualitative indicators by the expert group. S2.

22. Calculate the objective weight of evaluation indicators: Among them, h j is the objective weight vector of the evaluation index, e j is the entropy value of the jth indicator.

5. According to claim 1, a comprehensive evaluation method for distribution automation terminal status based on AKFCM model is characterized in that: The final combined weight is determined by game theory combined weighting: Among them, ω is the combination weight vector, w j is the weight vector of subjective weight, h j is the weight vector of objective weights, α1 is the combination coefficient of subjective weights, α2 is the combination coefficient of objective weights, β1 is the combination coefficient of subjective weights after game, and β2 is the combination coefficient of objective weights after game.

6. According to the AKFCM model-based comprehensive evaluation method for distribution automation terminal status according to claim 1, it is characterized in that: The method of using the multi-attribute decision-making method VIKOR to calculate the index value of the current clustering result and compare it with the ideal state for evaluation includes the following steps: S2.

31. Establish the initial evaluation matrix A: A = [a pq ]n×m, where q=(1,2,...,n), p=(1,2,...,m), m is the number of evaluation objects, n is the number of evaluation indicators, and a pq is the pth evaluation index value of the qth object to be evaluated; S2.

32. Establish a normalized matrix B and perform dimensionless processing on the initial evaluation matrix A: For benefit-oriented indicators, the bigger the better. For cost-type indicators, the smaller the better. Among them, b pq for a pq The dimensionless form of are the minimum and maximum values ​​of the qth column elements of the initial evaluation matrix A, respectively; Determine the positive ideal solution and negative ideal solution Count the team performance values ​​S of all objects to be evaluated p And personal regret value R: Calculate the benefit value Q of each object to be evaluated p : Among them, λ is the decision-making mechanism coefficient, λ∈[0,1], λ>0.5 means that the decision is made based on maximizing group benefits; λ<0.5 means that the decision is made based on minimizing individual regret.

Images