A distribution network performance index evaluation method based on subjective and objective combined weighting method

Abstract

The invention discloses a distribution network performance index evaluation method based on a subjective and objective combined weighting method, comprising: step 1, establishing a judgment matrix for determining a single sample weight of a power grid evaluation index according to a power grid evaluation standard; step 2, establishing a judgment matrix for determining a sample set weight of a power grid evaluation index; step 3, calculating the weights of each index of the two judgment matrices by using a genetic analytic hierarchy process AGA-CAHP for correcting the consistency of the judgment matrix; step 4, determining the combined weight of sample set data {x(j,i)} of n evaluation indicators of m samples; step 5, multiplying and accumulating the combined weight of the evaluation index value of each sample with the corresponding relative membership value to obtain a comprehensive index value of each sample; and solving the technical problems that the conventional evaluation method can no longer meet the new requirements in the face of new situations and new challenges in the prior art for power grid evaluation technology.

Description

Technical Field

[0001] The present invention belongs to the technical field of distribution network index evaluation, and in particular relates to a distribution network performance index evaluation method based on a subjective and objective combined weighting method. Background Art

[0002] Distribution network evaluation helps to discover construction problems and improvement potential of distribution networks, and is of great significance for guiding the planning and construction of renewable energy grid connection.

[0003] At present, the evaluation of distribution network is mainly the evaluation of each attribute of distribution network, and the evaluation of multiple indicators is less. At the same time, most of the multi-indicator evaluation methods are improved on the basis of analytic hierarchy process, mainly including:

[0004] The intelligent evaluation method based on subjective weighting is an evaluation method that weights each indicator according to the evaluator's subjective judgment on the evaluation indicator. Therefore, the weighting of indicators in this evaluation method can better reflect the evaluator's intention. At the same time, the subjective weighting method has a low requirement for the number of samples and a relatively simple solution. However, the rationality of the intelligent evaluation method based on subjective weighting is greatly affected by the evaluator and does not have sufficient mathematical basis.

[0005] The intelligent evaluation method based on objective weighting is an evaluation method that weights each evaluation object according to the objective role of each indicator in the evaluation process. Compared with the subjective weighting method, it does not take into account the evaluator's preferences and experience, but has a strong theoretical basis. At the same time, it also has high requirements on the quantity and quality of samples, so the calculation process is also more complicated.

[0006] The intelligent evaluation method based on similarity theory uses the method of simulated intelligence and similar analogy to summarize, perfect and improve the evaluation method used. It combines high abstraction with broad applicability, but its accuracy cannot be guaranteed when applied to more complex subjects.

[0007] The intelligent evaluation method based on pattern recognition is an evaluation method that uses pattern recognition method to determine the functional relationship between the evaluation index value and the comprehensive evaluation index. This method requires a large amount of data for identification.

[0008] The intelligent evaluation method based on decision analysis is an evaluation method that selects the best feasible option from several feasible options on the basis of certain information and experience. However, this method has uncertainty, and the decision criteria are often incommensurable and contradictory.

[0009] The intelligent evaluation method based on combination theory can propose different organizational evaluation methods according to different combination theories. It can integrate the advantages of multiple evaluation methods, but how to optimize the integration is a major problem.

[0010] Existing technologies are facing new situations and new challenges. Traditional evaluation methods can no longer meet new demands, so it is particularly important to establish new evaluation methods. Summary of the invention

[0011] The technical problem to be solved by the present invention is to provide a distribution network performance index evaluation method based on a subjective and objective combined weighting method to solve the technical problems that the existing technology faces new situations and new challenges for power grid evaluation technology, and traditional evaluation methods can no longer meet new needs.

[0012] Technical solution of the present invention:

[0013] A distribution network performance index evaluation method based on a subjective and objective combined weighting method, the method comprising:

[0014] Step 1: Establish a judgment matrix for determining the single sample weight of the power grid evaluation index according to the power grid evaluation standard;

[0015] Step 2: Establish a judgment matrix for determining the weights of the power grid evaluation index sample set;

[0016] Step 3, using the genetic analytic hierarchy process AGA-CAHP to modify the consistency of the judgment matrix to calculate the weight of each indicator of the two judgment matrices;

[0017] Step 4: Determine the combined weight of the sample set data {x(j,i)} of the n evaluation indicators of the m samples;

[0018] Step 5: Multiply the combined weight of each sample evaluation index value by the corresponding relative membership value and add them up to obtain the comprehensive index value of each sample.

[0019] The construction of the judgment matrix of the power grid evaluation index sample set weights includes: In order to eliminate the dimensional effect of each evaluation index and make the modeling universal, the sample data set {x * (j,i)} are unified and dimensionless, and the processing formula for the larger index with the larger positive impact on the power grid is taken as:

[0020]

[0021] The processing formula for the indicator with smaller value and greater positive impact on the power grid is:

[0022]

[0023] The processing formula for the index with a larger positive impact on the power grid is:

[0024]

[0025] In the formula, x p(i), x1(i), x mid (i) are the critical value of the highest level i-th evaluation indicator, the critical value of the lowest level i-th evaluation indicator and the intermediate optimal value of the i-th evaluation indicator in the performance indicator evaluation standard table, respectively; r(j,i) is the relative membership value of the i-th evaluation indicator of the j-th sample to the upper level indicator;

[0026] According to the value of s(i) of each evaluation index, we can get Judgment Matrix of Grade Judgment Scale

[0027]

[0028] Where s min 、s max are the minimum and maximum values ​​of {s(i)|i=1~n} respectively; the relative importance parameter value b m =min{9,s max / s min}, min is the smaller function.

[0029] The method of calculating the weight of each index of two judgment matrices by using the genetic analytic hierarchy process AGA-CAHP with modified judgment matrix consistency includes: step A, establishing a hierarchical structure model, step B, establishing judgment matrices of each layer, and step C, consistency inspection, modification and single ranking weight calculation of each judgment matrix.

[0030] The consistency check, correction and single ranking weight calculation methods of each judgment matrix include: assuming that the correction judgment matrix of B is X = {x ij} n×n The single ranking weight of each element is recorded as {w k |k=1~n}, then the X matrix that minimizes the following formula is the optimal consistency judgment matrix of B:

[0031]

[0032] stx ii =1(i=1~n)

[0033] 1 / x ji =x ij ∈[b ij -db ij ,b ij +db ij ]∩[1 / 9,9](i=1~n,j=i+1~n)

[0034] w k >0(k=1~n)

[0035]

[0036] In the formula, the objective function CIC(n) is called the consistency index coefficient, d is a non-negative parameter, the smaller the value CIC(n) on the left side of the formula, the higher the consistency of the judgment matrix B. When the global minimum value CIC(n)=0 is taken, X=B is established, and the judgment matrix B has complete consistency; when the CIC(n) value is less than a certain standard value, it is considered that the judgment matrix B has satisfactory consistency, and the single ranking weights of each element calculated based on this are acceptable; otherwise, the parameter d should be increased or B should be adjusted until satisfactory consistency is achieved. The consistency test, correction and single ranking calculation of each judgment matrix in the C layer can be obtained in the same way;

[0037] Then, the total ranking of the levels and its consistency test, i.e., determining the ranking weights of each element at the same level for the highest level, i.e., level A elements, and testing the consistency of each judgment matrix;

[0038] Finally, the optimal ranking of each decision plan is determined according to the total ranking weight of each factor in layer C.

[0039] The method for determining the combined weight of sample set data {x(j,i)} of n evaluation indicators of m samples includes: if the single indicator primary indicator of the sample {x(j,i)} value is k∈{1,2,…,p}, then the single sample weight of the sample x(j,i) value is w(j,i,1)=w(k,1), j=1~m, i=1~n; the sample set weight of the sample x(j,i) value is w(j,i,2)=w(i,2), i=1~n, j=1~m; then the sample set weight of the sample x(j,i) value is:

[0040] w(j,i)=αw(j,i,1)+(1-α)w(i,2) (5)

[0041] Where α∈[0,1] is the weight coefficient of the single sample weight.

[0042] The calculation of the comprehensive index value of each sample includes:

[0043]

[0044] The larger the comprehensive index value I(j), the more stable the jth sample is, and the samples are classified and sorted accordingly.

[0045] Beneficial effects of the present invention:

[0046] Improve the accuracy and scientificity of the evaluation: The subjective and objective combined weighting method combines the advantages of the subjective and objective weighting methods, and can comprehensively consider the subjective will of the decision maker and the objective information of the data. This method avoids the one-sidedness and limitations that may be brought about by a single weighting method, and improves the accuracy and scientificity of the evaluation. Through the subjective and objective combined weighting method, the weights of various performance indicators can be more reasonably allocated. This makes the evaluation results more in line with the actual situation and can more accurately reflect the performance status of the distribution network.

[0047] Improve improvement efficiency: Through the subjective and objective combined weighting method, the weak links and potential problems in the distribution network can be identified more accurately, making the improvement work more targeted and enabling faster optimization and upgrading of the distribution network.

[0048] Adaptability and flexibility: The subjective and objective combined weighting method can be flexibly adjusted according to the actual conditions of different regions and the characteristics of the distribution network, so that the method can be well applied in distribution networks of different regions and different sizes. At the same time, with the development and changes of the distribution network, the parameters and indicators of the subjective and objective combined weighting method can be easily adjusted and optimized, making the method have good scalability and adaptability, and can meet the needs of future distribution network performance evaluation.

[0049] To evaluate the pros and cons of different complex networks, it is necessary to obtain the weights corresponding to different networks. The analytic hierarchy process is a good and effective method for determining the weights of hierarchical structures. Therefore, the present invention uses the analytic hierarchy process to compare the evaluation indicators in pairs, thereby obtaining the weights. At the same time, it has certain advantages in combining qualitative and quantitative analysis of indicators.

[0050] It solves the technical problems that existing power grid evaluation technologies are facing new situations and new challenges, and traditional evaluation methods can no longer meet new demands. DETAILED DESCRIPTION

[0051] A distribution network performance index evaluation method based on a subjective and objective combined weighting method, wherein the distribution network performance index includes:

[0052] Vulnerability: The concept of grid vulnerability is proposed for the reliability and security of the grid. If the assessment and prevention are not done properly, it may cause a large-scale power outage and even affect the safe and stable operation of the entire grid system. Therefore, the safe and stable operation of the grid is closely related to the vulnerability of the grid. It is crucial to accurately identify and assess the vulnerability of the grid.

[0053] Robustness emphasizes the ability of the process to maintain system functions when the initial conditions change, while stability emphasizes the ability of the process to maintain system functions when the environment or the system itself changes.

[0054] The reliability of power system includes two aspects: redundancy and safety. The former refers to the fact that the power system has sufficient power generation capacity and transmission capacity in steady-state operation, and can meet the load requirements of power users under the conditions of meeting various operating constraints; the latter refers to the ability of the power system to maintain uninterrupted power supply without system loss of control when subjected to sudden disturbances.

[0055] Grid resilience refers to the ability of the power network to maintain its basic functions and recover quickly after being disturbed by external extreme events or malicious attacks. Grid resilience vividly reflects the system's ability to respond quickly to disturbances. Its most important characteristics are the elastic tolerance of the grid topology, elastic absorption under disturbances, response and recovery. Grid resilience can better cope with extreme events with low probability and high losses, minimize the impact and scope of the event, and have the ability to quickly restore power supply.

[0056] The evaluation method includes:

[0057] Step 1: According to the power grid evaluation standard, establish a judgment matrix A for determining the single sample weight of the power grid evaluation index. The single sample weight of each power grid evaluation index refers to the weight of the single index power grid evaluation corresponding to the index value. The larger the index value, the greater the impact on the power grid evaluation result. Based on this, the weight of the performance index calculated by AHP can be obtained. The judgment matrix of the level judgment scale A=(a ij ) p×p .

[0058] Step 2: Establish a judgment matrix B for determining the weights of the power grid evaluation index sample set. In order to eliminate the dimensional effect of each evaluation index and make the modeling universal, it is necessary to * (j,i)} are unified and dimensionless. The processing formula for the larger the value, the greater the positive impact on the power grid can be taken as

[0059]

[0060] The processing formula for the index with smaller value and greater positive impact on the power grid can be taken as

[0061]

[0062] The processing formula for the index with a larger positive impact on the power grid can be taken as

[0063]

[0064] In the formula, x p (i), x1(i), x mid(i) are the critical value of the highest level i-th evaluation indicator, the critical value of the lowest level i-th evaluation indicator and the intermediate optimal value of the i-th evaluation indicator in the performance indicator evaluation standard table. r(j,i) is the relative membership value of the i-th evaluation indicator of the j-th sample to the upper level indicator.

[0065] According to the value of s(i) of each evaluation index, we can get Judgment matrix of the level judgment scale.

[0066]

[0067] Where s min 、s max are the minimum and maximum values ​​of {s(i)|i=1~n} respectively; the relative importance parameter value b m =min{9,s max / s min}, min is the smaller function.

[0068] Step 3: Use the genetic analytic hierarchy process (AGA-CAHP) with modified judgment matrix consistency to calculate the weights of each indicator in the judgment matrix, which are recorded as single sample weight and sample set weight respectively.

[0069] The principle of the genetic hierarchical analysis method for correcting the consistency of the judgment matrix includes: step 1 establishing a hierarchical structure model, step 2 establishing the judgment matrix of each layer, and main step 3 consistency test, correction and single ranking weight calculation of each judgment matrix.

[0070] The AGA algorithm is applied in the consistency check, correction and single ranking weight calculation of each matrix. The specific content of step 3 is: let the correction judgment matrix of B be X = {x ij} n×n The single ranking weight of each element is still recorded as {w k |k=1~n}, then the X matrix that minimizes the following formula is called the optimal consistency judgment matrix of B:

[0071]

[0072] stx ii =1(i=1~n)

[0073] 1 / x ji =x ij ∈[b ij -db ij ,b ij +db ij ]∩[1 / 9,9](i=1~n,j=i+1~n)

[0074] w k>0(k=1~n)

[0075]

[0076] In the formula, the objective function CIC(n) is called the consistency index coefficient; d is a non-negative parameter, which can be selected from [0,0.5] according to experience; the other symbols are the same as before. Obviously, the smaller the value CIC(n) on the left side of the formula, the higher the consistency of the judgment matrix B. When the global minimum value CIC(n) = 0 is taken, X = B is established, and the judgment matrix B has complete consistency. The accelerated genetic algorithm (AGA) that simulates the biological survival of the fittest rule and the chromosome information exchange mechanism within the population is a general global optimization method, which is relatively simple and effective to solve. When the CIC(n) value is less than a certain standard value, it can be considered that the judgment matrix B has satisfactory consistency, and the single ranking weights of each element calculated based on this are acceptable; otherwise, it is necessary to increase the parameter d or adjust B until it has satisfactory consistency. The consistency test, correction and single ranking calculation of each judgment matrix in the C layer can be obtained in the same way.

[0077] Subsequently, the total hierarchical ranking and its consistency test, that is, determining the ranking weights of each element at the same level for the highest level (A level) elements and testing the consistency of each judgment matrix, is carried out layer by layer from the highest level to the lowest level.

[0078] Finally, the optimal ranking of each decision plan is determined according to the total ranking weight of each factor in layer C.

[0079] Step 4: Determine the combined weight {w(j,i)} of the sample set data {x(j,i)} of the n evaluation indicators of the m samples, j=1~m, i=1~n. If the single-indicator first-level indicator of the sample {x(j,i)} value is k∈{1,2,…,p}, then the single sample weight of the sample x(j,i) value is w(j,i,1)=w(k,1), j=1~m, i=1~n. The sample set weight of the sample x(j,i) value is w(j,i,2)=w(i,2), i=1~n, j=1~m. Then the sample set weight of the sample x(j,i) value is:

[0080] w(j,i)=αw(j,i,1)+(1-α)w(i,2) (5)

[0081] In the formula, α∈[0,1] is the weight coefficient of the single sample weight. Formula (5) shows that the combined weight determines its comprehensive weight according to the primary indicator and evaluation indicator type to which the sample evaluation index value belongs, and the sample information is more comprehensive.

[0082] Step 5: Multiply and add the combined weight w(j,i) of each sample evaluation index value with the corresponding relative membership value r(j,i) to obtain the comprehensive index value of each sample.

[0083]

[0084] The larger the comprehensive index value I(j), the more stable the jth sample is, and the samples can be classified and sorted based on this.

Claims

1. A distribution network performance index evaluation method based on subjective and objective combined weighting method, characterized by: The method comprises: Step 1: Establish a judgment matrix for determining the single sample weight of the power grid evaluation index according to the power grid evaluation standard; Step 2: Establish a judgment matrix for determining the weights of the power grid evaluation index sample set; Step 3, using the genetic analytic hierarchy process AGA-CAHP to modify the consistency of the judgment matrix to calculate the weight of each indicator of the two judgment matrices; Step 4: Determine the combined weight of the sample set data {x(j,i)} of the n evaluation indicators of the m samples; Step 5: Multiply the combined weight of each sample evaluation index value by the corresponding relative membership value and add them up to obtain the comprehensive index value of each sample.

2. The method for evaluating distribution network performance indicators based on a subjective and objective combined weighting method according to claim 1 is characterized in that: The construction of the judgment matrix of the power grid evaluation index sample set weights includes: In order to eliminate the dimensional effect of each evaluation index and make the modeling universal, the sample data set {x * (j,i)} are unified and dimensionless, and the processing formula for the larger index with the larger positive impact on the power grid is taken as: The processing formula for the indicator with smaller value and greater positive impact on the power grid is: The processing formula for the index with a larger positive impact on the power grid is: In the formula, x p (i), x1(i), x mid (i) are the critical value of the highest level i-th evaluation indicator, the critical value of the lowest level i-th evaluation indicator and the intermediate optimal value of the i-th evaluation indicator in the performance indicator evaluation standard table, respectively; r(j,i) is the relative membership value of the i-th evaluation indicator of the j-th sample to the upper level indicator; According to the value of s(i) of each evaluation index, we can get Judgment Matrix of Grade Judgment Scale Where s min 、s max are the minimum and maximum values ​​of {s(i)|i=1~n} respectively; the relative importance parameter value b m =min{9,s max / s min }, min is the smaller function.

3. The method for evaluating distribution network performance indicators based on a subjective and objective combined weighting method according to claim 1 is characterized in that: The method of calculating the weight of each index of two judgment matrices by using the genetic analytic hierarchy process AGA-CAHP with modified judgment matrix consistency includes: step A, establishing a hierarchical structure model, step B, establishing judgment matrices of each layer, and step C, consistency inspection, modification and single ranking weight calculation of each judgment matrix.

4. The method for evaluating distribution network performance index based on subjective and objective combined weighting method according to claim 3 is characterized in that: The consistency check, correction and single ranking weight calculation methods of each judgment matrix include: assuming that the correction judgment matrix of B is X = {x ij } n×n The single ranking weight of each element is recorded as {w k |k=1~n}, then the X matrix that minimizes the following formula is the optimal consistency judgment matrix of B: s.t.x ii =1(i=1~n) 1 / x ji =x ij ∈[b ij -db ij ,b ij +db ij ]∩[1 / 9,9](i=1~n,j=i+1~n) w k >0(k=1~n) In the formula, the objective function CIC(n) is called the consistency index coefficient, d is a non-negative parameter, the smaller the value CIC(n) on the left side of the formula, the higher the consistency of the judgment matrix B. When the global minimum value CIC(n)=0 is taken, X=B is established, and the judgment matrix B has complete consistency; when the CIC(n) value is less than a certain standard value, it is considered that the judgment matrix B has satisfactory consistency, and the single ranking weights of each element calculated based on this are acceptable; otherwise, the parameter d should be increased or B should be adjusted until satisfactory consistency is achieved. The consistency test, correction and single ranking calculation of each judgment matrix in the C layer can be obtained in the same way; Then, the total ranking of the levels and its consistency test, i.e., determining the ranking weights of each element at the same level for the highest level, i.e., level A elements, and testing the consistency of each judgment matrix; Finally, the optimal ranking of each decision plan is determined according to the total ranking weight of each factor in layer C.

5. The method for evaluating distribution network performance index based on subjective and objective combined weighting method according to claim 1 is characterized in that: The method for determining the combined weight of sample set data {x(j,i)} of n evaluation indicators of m samples includes: if the single indicator primary indicator of the sample {x(j,i)} value is k∈{1,2,…,p}, then the single sample weight of the sample x(j,i) value is w(j,i,1)=w(k,1), j=1~m, i=1~n; the sample set weight of the sample x(j,i) value is w(j,i,2)=w(i,2), i=1~n, j=1~m; then the sample set weight of the sample x(j,i) value is: w(j,i)=αw(j,i,1)+(1-α)w(i,2) (5) Where α∈[0,1] is the weight coefficient of the single sample weight.

6. The method for evaluating distribution network performance index based on subjective and objective combined weighting method according to claim 1 is characterized in that: The calculation of the comprehensive index value of each sample includes: The larger the comprehensive index value I(j), the more stable the jth sample is, and the samples are classified and sorted accordingly.

Images