A method for evaluating the state of power utilization in a hydropower station based on ANP and grey correlation analysis
By combining ANP and grey relational analysis, a multi-dimensional indicator system for plant power systems is constructed. The indicator weights are quantified and the correlation degree is dynamically calculated, which solves the problems of unreasonable weight allocation and strong subjectivity in traditional methods, and realizes accurate status assessment and efficient decision-making for plant power systems.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHINA THREE GORGES UNIV
- Filing Date
- 2026-03-31
- Publication Date
- 2026-07-07
AI Technical Summary
Existing technologies struggle to handle the complex dependencies and feedback relationships among evaluation indicators of plant power systems. Traditional methods rely on subjective weight settings and lack dynamic correlation quantification, resulting in insufficient objectivity in evaluation results and an inability to achieve accurate evaluation.
A multi-dimensional indicator system is constructed by combining ANP and grey relational analysis. The weights of the indicators are quantified by ANP network hierarchical analysis, and the correlation degree is dynamically calculated by grey relational analysis. Combined with ANP weighted correction, the accurate status assessment of the plant power system is achieved.
By scientifically quantifying the weights of indicators, we can enhance the adaptability of data and the robustness of assessment, achieve accurate state classification and efficient decision-making, improve the accuracy of assessment results by more than 30%, enhance the anti-interference ability by 40%, and achieve a risk identification accuracy rate of 92%.
Smart Images
Figure CN122347352A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of power system condition assessment technology, and in particular to a method for assessing the condition of hydropower plant power consumption based on ANP and grey relational analysis. Background Technology
[0002] In the field of power system operation and maintenance, the power plant auxiliary power system, as the core link of power plant internal power supply, is directly related to the safe and stable operation and economic benefits of generator units through accurate assessment of its operating status. With the advancement of smart grid construction, the power plant auxiliary power system is showing a trend of increasingly complex equipment interconnection, diversified operating data, and diversified fault causes. Traditional status assessment methods are no longer sufficient to meet the actual needs of multi-dimensional index coupling relationship analysis and uncertain information processing. In existing technologies, the evaluation index system of the power plant auxiliary power system has significant network characteristics, with complex dependencies and feedback relationships between various indicators (such as power supply reliability, load balance, and protection device coordination). Traditional methods are unable to accurately characterize this nonlinear correlation. At the same time, traditional evaluation methods are highly dependent on subjective weight setting and lack quantitative analysis of the dynamic correlation between indicators, resulting in insufficient objectivity and robustness of the evaluation results. In addition, when dealing with multi-criteria decision-making problems, existing technologies generally suffer from defects such as strong subjectivity in index weight assignment and weak adaptability of evaluation models to complex systems, making it difficult to achieve refined diagnosis of the operating status of the power plant auxiliary power system.
[0003] Therefore, the applicant proposes a method for assessing the power status of hydropower plants based on ANP and grey relational analysis. Summary of the Invention
[0004] The purpose of this invention is to solve the following technical problems: existing analytic hierarchy process (AHP) methods are difficult to handle the complex dependencies and feedback relationships between evaluation indicators of plant power systems, resulting in unreasonable weight allocation; existing evaluation methods rely on subjective weight setting, lack dynamic correlation quantitative analysis, and have insufficient objectivity in evaluation results; existing technologies cannot take into account the network relationships of indicators, uncertain data characteristics, and dynamic weight calculation, making it difficult to achieve accurate evaluation of the operating status of plant power systems; and this invention is proposed to address these issues.
[0005] To solve the above-mentioned technical problems, the present invention proposes the following technical solution: A method for assessing the power status of hydropower plants based on ANP and grey relational analysis includes the following steps: S1: Construct a multi-dimensional indicator system for power plant power consumption in hydropower stations; S2: Obtain the global weights of the status assessment indicators for the power plant's auxiliary power system; S3: Collect relevant data on the power consumption indicators of hydropower plants, and perform data standardization and processing; S4: Construct the correlation coefficient matrix, calculate the grey correlation coefficient, and continue to calculate the weighted grey correlation degree of the fused ANP weights; S5: Determine the safety level of the final target layer, that is, obtain the final evaluation result of the plant power system.
[0006] In step S1, starting from the equipment characteristics, operating environment, and management requirements of the hydropower station's power supply system, an indicator system covering multiple dimensions such as equipment status, operation and maintenance, safety risks, and automation level is constructed.
[0007] In step S2, the ANP weighting system is constructed, including the following sub-steps: 2-1) Construct the ANP influence matrix; 2-2) Construct the ANP judgment matrix; 2-3) Construct the hypermatrix; 2-4) Construct a weighted hypermatrix; 2-5) Construct the limit supermatrix and obtain the global weight of the ANP index.
[0008] In 2-1), hydropower station operation and maintenance experts and safety management personnel were invited to establish an ANP influence relationship matrix for the comprehensive evaluation index system of hydropower station power supply. The rows and columns of the matrix are all power supply status evaluation indicators. In the matrix, "1" indicates that the corresponding index in the row has an impact on the index in the column, and "0" indicates no impact.
[0009] In 2-2), the ANP network layer indicators are scored according to the indirect dominance principle to construct a judgment matrix. The specific method is as follows: Suppose that the control layer of the ANP network structure of the power plant auxiliary power system of a hydropower station has m indices B1, B2, B3, ..., Bm, and the network layer has N indices C1, C2, C3, ..., Cn. Using a certain index Ci (i=1, 2, ..., n) of the network layer as a criterion, the importance of all indices affecting Ci under the corresponding control layer Bm is compared pairwise. The nine-fold scaling method is used to construct the ANP judgment matrix for Ci under the control layer Bm. By analogy, the ANP judgment matrices of all control layers Bm and network layers Ci are constructed.
[0010] In 2-3), construct the supermatrix. The specific steps are as follows: First, the eigenvectors of the normalized judgment matrix are calculated using the eigenvalue method, which are the weight vectors of the judgment matrix. After the consistency check passes, the normalized eigenvectors obtained from the judgment matrices of all indicators within the control layer Bm are combined to obtain the weight vector matrix. : ; in, The column vector in the text is a normalized feature vector calculated by using a certain indicator in the multi-dimensional indicator system Bm of hydropower plant power consumption as a secondary criterion and constructing a judgment matrix using the indicators in Bi that affect that indicator. Formula for the consistency test coefficient (CR number): ; By default, a CR < 0.1 is sufficient to pass the consistency test. Here, n represents the order of each judgment matrix constructed for the hydropower station's power supply system. The largest eigenvalue of the judgment matrix in this verification. Secondly, repeat the above operation for all metrics within all network layers C, and integrate all calculated weight vector matrices to obtain a supermatrix. : ; Among them, supermatrix Each column is a combination of weight vectors representing the influence of indicators on a certain indicator in the multi-dimensional indicator system network layer of all hydropower plant power consumption, where n is the total number of indicator elements in the network layer.
[0011] In section 2-4), the specific steps for constructing the weighted hypermatrix are as follows: First, construct the weighted matrix A. Specifically, calculate the normalized eigenvectors of the control layer's judgment matrix and arrange and integrate them to obtain the weighted matrix A. ; Among them, the first column of the weighting matrix A is the weight vector of all control layer indicators of the multi-dimensional indicator system for power plant power of all hydropower stations, when the first control layer indicator is the secondary criterion, and so on. Secondly, regarding the supermatrix in steps 2-3) Weighting is performed using a weighting matrix A, that is, matrix A and the hypermatrix. Block multiplication yields the weighted hypermatrix : .
[0012] Among them, the weighted supermatrix This reflects the direct influence relationship between multiple indicators of power plant power consumption in hydropower stations and the importance weight between groups, where n is the total number of all indicator elements in the network layer.
[0013] In section 2-5), the specific steps for constructing the limiting hypermatrix and obtaining the global weights of ANP are as follows: when When the limit hypermatrix is obtained, Multiplying itself to infinite powers until the matrix result converges, the column vectors of the limiting hypermatrix at this point are the global weights of the hydropower plant power system state assessment index obtained using the ANP method.
[0014] In step S3, based on the established multi-dimensional indicator system, specifically, data is collected in real time through multiple channels such as SCADA system, smart sensors, and equipment inspection terminals. Since the collected hydropower plant power data has different dimensions and positive and negative polarities, it needs to be standardized and preprocessed. Positive performance indicators for hydropower plant power, such as insulation resistance and protection operation accuracy, are processed using the following formulas: ; Negative power output indicators for hydropower plants, such as voltage deviation, harmonic distortion rate, and temperature rise, are specifically processed using the following formulas: ; in, This represents the original value of the k-th indicator for the i-th sample of the hydropower station's power supply system. , This is the standardized value.
[0015] In step S4, a correlation coefficient matrix is constructed using grey relational analysis, thereby calculating the grey relational coefficients; the specific steps are as follows: 4-1) Let the standard level be... There are a total of 5 safety levels, and a standard reference sequence for various indicators of power plant power consumption data of hydropower stations is established; ; Among them, the above matrix Each row represents a status level, and each column represents the standard values of each level of power plant power consumption indicators after standardization.
[0016] Set the set of sequences to be analyzed for association with the reference sequences as comparison sequences. That is, the real-time data of the corresponding indicators of the power plant's power consumption obtained from the collection, consisting of m comparison sequences, each sequence having n values.
[0017] 4-2) Calculate the grey relational coefficient between the actual data and standard data of the hydropower station's power supply system; According to the grey system theory, a grey system is a system in which some information is known and some information is unknown, and it is used to solve the uncertainty caused by incomplete information. The status assessment of the power plant system is assumed to be a grey system, and the safety assessment of the power plant system can be carried out through grey relational analysis (GRA). ; in, Let be the grey coefficient between the k-th indicator and the k-th optimal indicator of the i-th evaluation object in the plant power status evaluation. It is the gray coordination coefficient, usually ; ; ; This represents the absolute difference between the reference sequence and the comparison sequence at each time point. The minimum value of the difference sequence. Maximum value of the difference sequence 。
[0018] In step S5, the grey relational coefficient obtained by integrating the weighted grey relational degree of ANP is calculated. By constructing a relational degree calculation formula with ANP weight correction factor, the correlation degree between the actual operation data of the hydropower plant power system and the standard reference data is weighted and corrected. Finally, a weighted grey relational degree evaluation value that can comprehensively reflect the difference in the weight of the indicator network is formed, and the state level of the final target layer is determined, that is, the final evaluation result of the hydropower plant power system is given. Weighted grey relational degree calculation formula: ; in, The global weight of the k-th index calculated for the plant power system ANP. The first step in the condition assessment of the plant power system Grey weighted correlation between each evaluation indicator and the ideal result; ; For the status evaluation of plant power consumption, the first i For each evaluation object, calculate its correlation with five standard levels. 1, 2, 3, 4, 5. Select the level with the highest correlation as the power supply operation status level for this sample: .
[0019] Compared with the prior art, the present invention has the following technical effects: 1) Scientific Quantification of Indicator Weights: This invention uses the ANP network analytic hierarchy process to construct an index network model, which solves the problem that the traditional analytic hierarchy process cannot characterize the dependency and feedback relationships of indicators. It quantifies the importance of the indicators themselves and the feedback weights between groups, making the weight allocation more in line with the actual characteristics of the power system. The objectivity of the index weights is improved by more than 30% compared with the traditional method, providing a reliable quantitative basis for evaluation.
[0020] (ii) Enhance data adaptability and assessment robustness; This invention addresses gray features such as missing data and noise by introducing gray relational analysis to dynamically calculate the correlation degree, combined with ANP differential weighting correction to reduce subjective influence. Even in scenarios where data integrity is below 70%, it can still effectively extract information, and its anti-interference capability is 40% higher than traditional fuzzy evaluation methods, making it suitable for state diagnosis in environments with strong electromagnetic interference.
[0021] (iii) Achieving accurate state classification and efficient decision-making; This invention integrates ANP weight calculation and grey relational analysis to construct a full-process evaluation method. Through a multi-level state reference sequence library and maximum correlation rules, it accurately identifies risk points such as insufficient power capacity and load imbalance. Practical applications show that the evaluation cycle is shortened by 50%, and the risk identification accuracy rate reaches over 92%, providing efficient and reliable support for operation and maintenance decisions. Attached Figure Description
[0022] The present invention will be further described below with reference to the accompanying drawings and embodiments: Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is a diagram of the plant power system status assessment index system based on the ANP method in this invention; Figure 3 This is a histogram showing the safety level of the power supply system for a hydropower station as described in this invention. Detailed Implementation
[0023] like Figure 1 As shown, a method for assessing the power status of a hydropower station based on ANP and grey relational analysis includes the following steps: S1: Construct a multi-dimensional indicator system for power plant power consumption in hydropower stations; S2: The global weights of the hydropower station's power system status assessment indicators are calculated using the ANP network hierarchical analysis method. S3: Collect relevant data on the power consumption indicators of hydropower plants, and perform data standardization and processing; S4: Construct a correlation coefficient matrix using grey relational analysis, calculate the grey relational coefficients, and then calculate the weighted grey relational degree fused with ANP weights. S5: Determine the safety level of the final target layer, that is, obtain the final evaluation result of the plant power system.
[0024] In step S1, starting from the equipment characteristics, operating environment, and management requirements of the hydropower station's power supply system, an indicator system covering multiple dimensions such as equipment status, operation and maintenance, safety risks, and automation level is constructed.
[0025] In step S2, the ANP weighting system is constructed, including the following sub-steps: 2-1) Construct the ANP influence matrix; 2-2) Construct the ANP judgment matrix; 2-3) Construct the hypermatrix; 2-4) Construct a weighted hypermatrix; 2-5) Construct the limit supermatrix and obtain the global weight of the ANP index.
[0026] In section 2-1), hydropower station operation and maintenance experts and safety management personnel were invited to establish an ANP influence relationship matrix for the comprehensive evaluation index system of hydropower station plant power, as shown in Table 1 of the embodiment. In this matrix, the rows and columns are the plant power status evaluation indicators set in the embodiment. In Table 1, "1" indicates that the corresponding indicator in the row has an impact on the indicator in the column, and "0" indicates no impact.
[0027] In 2-2), the ANP network layer indicators are scored according to the indirect dominance principle to construct a judgment matrix. The specific method is as follows: Suppose that the control layer of the ANP network structure of the hydropower station's power supply system has m indices B1, B2, B3, ..., Bm, and the network layer has N indices C1, C2, C3, ..., Cn. Using a certain index Ci (i = 1, 2, ..., n) in the network layer as a criterion, the importance of all indices affecting Ci under the corresponding control layer Bm is compared pairwise. The nine-fold scaling method is used to construct the ANP judgment matrix for Ci under the control layer Bm. This process is repeated to construct the ANP judgment matrices for all control layers Bm and network layers Ci. The established judgment matrices are shown in Tables 2, 3, and 4 in this embodiment.
[0028] In 2-3), construct the supermatrix. The specific steps are as follows: First, the eigenvectors of the normalized judgment matrix are calculated using the eigenvalue method, which are the weight vectors of the judgment matrix. After the consistency check passes, the normalized eigenvectors obtained from the judgment matrices of all indicators within the control layer Bm are combined to obtain the weight vector matrix. : ; in, The column vector in the text is a normalized feature vector calculated by using a certain indicator in the multi-dimensional indicator system Bm of hydropower plant power consumption as a secondary criterion and constructing a judgment matrix using the indicators in Bi that affect that indicator. Formula for the consistency test coefficient (CR number): ; By default, a CR < 0.1 is sufficient to pass the consistency test. Here, n represents the order of each judgment matrix constructed for the hydropower station's power supply system. The largest eigenvalue of the judgment matrix in this verification. Secondly, repeat the above operation for all metrics within all network layers C, and integrate all calculated weight vector matrices to obtain a supermatrix. : ; Among them, supermatrix Each column is a combination of weight vectors representing the influence of indicators on a certain indicator in the multi-dimensional indicator system network layer of all hydropower plant power consumption, where n is the total number of indicator elements in the network layer.
[0029] In section 2-4), the specific steps for constructing the weighted hypermatrix are as follows: First, construct the weighted matrix A. Specifically, calculate the normalized eigenvectors of the control layer's judgment matrix and arrange and integrate them to obtain the weighted matrix A. ; Among them, the first column of the weighting matrix A is the weight vector of all control layer indicators of the multi-dimensional indicator system for power plant power of all hydropower stations, when the first control layer indicator is the secondary criterion, and so on. Secondly, regarding the supermatrix in steps 2-3) Weighting is performed using a weighting matrix A, that is, matrix A and the hypermatrix. Block multiplication yields the weighted hypermatrix : .
[0030] Among them, the weighted supermatrix This reflects the direct influence relationship between multiple indicators of power plant power consumption in hydropower stations and the importance weight between groups, where n is the total number of all indicator elements in the network layer.
[0031] In section 2-5), the specific steps for constructing the limiting hypermatrix and obtaining the global weights of ANP are as follows: when When the limit hypermatrix is obtained, Multiplying itself to infinite powers until the matrix result converges, the column vectors of the limiting hypermatrix at this point are the global weights of the hydropower plant power system state assessment index obtained using the ANP method.
[0032] In step S3, based on the established multi-dimensional indicator system, specifically, data is collected in real time through multiple channels such as SCADA system, smart sensors, and equipment inspection terminals. Since the collected hydropower plant power data has different dimensions and positive and negative polarities, it needs to be standardized and preprocessed. Positive performance indicators for hydropower plant power, such as insulation resistance and protection operation accuracy, are processed using the following formulas: ; Negative power output indicators for hydropower plants, such as voltage deviation, harmonic distortion rate, and temperature rise, are specifically processed using the following formulas: ; in, This represents the original value of the k-th indicator for the i-th sample of the hydropower station's power supply system. , This is the standardized value.
[0033] 10. The method according to claim 1, characterized in that, in step S4, a correlation coefficient matrix is constructed by grey relational analysis, thereby calculating the grey relational coefficient; the specific steps are as follows: 4-1) Let the standard level be... There are a total of 5 safety levels, and a standard reference sequence for various indicators of power plant power consumption data of hydropower stations is established; ; Among them, the above matrix Each row represents a status level, and each column represents the standard values of each level of power plant power consumption indicators after standardization.
[0034] Set the set of sequences to be analyzed for association with the reference sequences as comparison sequences. That is, the real-time data of the corresponding indicators of the power plant's power consumption obtained from the collection, consisting of m comparison sequences, each sequence having n values.
[0035] 4-2) Calculate the grey relational coefficient between the actual data and standard data of the hydropower station's power supply system; According to the grey system theory, a grey system is a system in which some information is known and some information is unknown, and it is used to solve the uncertainty caused by incomplete information. The status assessment of the power plant system is assumed to be a grey system, and the safety assessment of the power plant system can be carried out through grey relational analysis (GRA). ; in, Let be the grey coefficient between the k-th indicator and the k-th optimal indicator of the i-th evaluation object in the plant power status evaluation. It is the gray coordination coefficient, usually ; ; ; This represents the absolute difference between the reference sequence and the comparison sequence at each time point. The minimum value of the difference sequence. Maximum value of the difference sequence 。
[0036] In step S5, the grey relational coefficient obtained by integrating the weighted grey relational degree of ANP is calculated. By constructing a relational degree calculation formula with ANP weight correction factor, the correlation degree between the actual operation data of the hydropower plant power system and the standard reference data is weighted and corrected. Finally, a weighted grey relational degree evaluation value that can comprehensively reflect the difference in the weight of the indicator network is formed, and the state level of the final target layer is determined, that is, the final evaluation result of the hydropower plant power system is given. Weighted grey relational degree calculation formula: ; in, The global weight of the k-th index calculated for the plant power system ANP. The first step in the condition assessment of the plant power system Grey weighted correlation between each evaluation indicator and the ideal result; ; For the status evaluation of plant power consumption, the first i For each evaluation object, calculate its correlation with five standard levels. 1, 2, 3, 4, 5. Select the level with the highest correlation as the power supply operation status level for this sample: .
[0037] Example: A method for assessing the state of power consumption in a plant based on ANP network hierarchical analysis and grey relational analysis includes: S1: Based on the system characteristics of the hydropower station's power supply system, construct an objective and reasonable multi-dimensional indicator system for the power supply of hydropower stations. S2: After establishing an objective and reasonable evaluation index system for the power plant's auxiliary power system, in order to quantify the overall goal of the power plant's auxiliary power system status evaluation, the global weights of the power plant's auxiliary power system status evaluation indexes are calculated using the ANP network hierarchical analysis method. S3: After determining the weight of each indicator, start collecting relevant data on the power consumption indicators of the hydropower station and perform data standardization and processing. S4: After data standardization and processing, construct a correlation coefficient matrix using grey relational analysis, calculate the grey relational coefficients, and then calculate the weighted grey relational degree using the fused ANP weights. S5: Determine the safety level of the final target layer, that is, obtain the final evaluation result of the plant power system.
[0038] 1. First, based on the equipment characteristics, operating environment, and management requirements of the hydropower station's power supply system, a multi-dimensional indicator system is constructed, encompassing equipment status, operation and maintenance, safety risks, and automation levels. For example... Figure 2 As shown.
[0039] 2. Establish an ANP network structure, and have experts in the field of hydropower determine the elements and the influence relationships between element groups under each level of evaluation indicators of the plant power system, and construct the influence relationship matrix as shown in Table 1 below;
[0040] In Table 1, A1-A4 represent network layer indicators: motor vibration amplitude, bearing temperature, main transformer winding temperature, and partial discharge quantity, respectively; B1-B3 represent network layer indicators: power consumption of the speed regulating oil pump, power consumption of the cooling water pump, and power consumption of the air compressor, respectively; C1-C3 represent network layer indicators: fault frequency, emergency response time, and number of protection actions, respectively; D1-D3 represent network layer indicators: busbar connection method, short-circuit current magnitude, and system redundancy, respectively. The influence relationships are specifically indicated from row to column, with 1 indicating an influence relationship between two elements and 0 indicating no influence relationship.
[0041] 3. Establish ANP judgment matrices. Based on the influence relationship matrix, construct network-level judgment matrices and control-level judgment matrices respectively; Taking A1 as an example in Table 2, from Table 1, Figure 1 It can be concluded that among the equipment status performance indicators in the control layer, only A2 and A4 affect the motor vibration amplitude A1 index. Therefore, based on the importance of A2 and A4 to A1, experts were invited to use the 9-fold scaling method shown in Table 3 to score, and the judgment matrix Table 2 was obtained.
[0042]
[0043] Similarly, for each network layer indicator, a judgment matrix of other factors affecting the indicator is constructed in the network layers under each control layer. Due to space limitations, they will not be shown one by one. For each control layer indicator, a judgment matrix of other control layers affecting the control layer is constructed as shown in Table 4.
[0044] The normalized eigenvectors are calculated using the eigenvalue method. After the consistency check passes, the normalized eigenvectors obtained from the judgment matrices constructed under Bi for all indicators in Bi are combined, and all judgment matrices are subjected to a consistency CR check. After the consistency check passes, the normalized eigenvectors obtained from the judgment matrices are combined to obtain the weight vector matrix.
[0045] 4. Construct the unweighted hypermatrix of ANP. Based on the original hypermatrix constructed from the judgment matrix, i.e. the control layer weight matrix, the unweighted hypermatrix of each index of the network layer is constructed as shown in Table 5;
[0046] 5. Establish the ANP weighted supermatrix. Using index Bm under the control layer of the plant power consumption index system as the secondary criterion, the importance of the control layer indices is compared pairwise to obtain the judgment matrix, and the normalized vector is calculated. The weighted matrix obtained from this calculation is shown in Table 6, which is the control layer weight matrix of the plant power consumption system in this example.
[0047] Multiply the weight matrix in Table 6 with the unweighted supermatrix of each network layer index in Table 5 in blocks to obtain the ANP weighted supermatrix as shown in Table 7.
[0048] 6. Establish the ANP limiting hypermatrix and obtain the ANP global weights. Table 8 shows the limiting hypermatrix of each index of the plant power system network layer obtained using the method described in this claim, and Table 9 shows the state assessment weight results of the plant power system under the ANP method obtained using the method described in this claim.
[0049]
[0050]
[0051] 7. Based on the established multi-dimensional indicator system of the plant power system, specifically, data is collected in real time through multiple sources such as SCADA system, smart sensors, and equipment inspection terminals. The collected hydropower plant power data is standardized and preprocessed to address the differences in dimensions and positive and negative polarities.
[0052] 8. Construct a correlation coefficient matrix using grey relational analysis to calculate the grey relational coefficients. Let the standard level be... Five safety levels; the set of sequences to be analyzed for correlation with the reference sequence is set as the comparison sequence; based on grey system theory, the grey correlation coefficient between the actual data and the standard data of the plant power system is calculated.
[0053] 9. Calculate the grey relational coefficient obtained by integrating the ANP weighted grey relational degree. By constructing a relational degree calculation formula containing the ANP weight correction factor, a weighted grey relational degree evaluation value that can comprehensively reflect the differences in the weights of the indicator network is finally formed, and the final evaluation result of the plant power system is obtained as follows: Figure 3 As shown.
Claims
1. A method for assessing the power status of hydropower plants based on ANP and grey relational analysis, characterized in that, Includes the following steps: Step S1: Construct a multi-dimensional indicator system for power plant power consumption in hydropower stations; Step S2: Obtain the global weights of the status assessment indicators of the hydropower station's power supply system; Step S3: Collect relevant data on the power consumption indicators of the hydropower station and perform data standardization and processing; Step S4: Construct the correlation coefficient matrix, calculate the grey correlation coefficient, and continue to calculate the weighted grey correlation degree of the fused ANP weights; Step S5: Determine the safety level of the final target layer, that is, obtain the final evaluation result of the plant power system.
2. The method according to claim 1, characterized in that, In step S1, starting from the equipment characteristics, operating environment, and management requirements of the hydropower station's power supply system, an indicator system covering multiple dimensions such as equipment status, operation and maintenance, safety risks, and automation level is constructed.
3. The method according to claim 1, characterized in that, In step S2, the ANP weighting system is constructed, including the following sub-steps: 2-1) Construct the ANP influence matrix; 2-2) Construct the ANP judgment matrix; 2-3) Construct the hypermatrix; 2-4) Construct a weighted hypermatrix; 2-5) Construct the limit supermatrix and obtain the global weight of the ANP index.
4. The method according to claim 3, characterized in that, In 2-1), hydropower station operation and maintenance experts and safety management personnel were invited to establish an ANP influence relationship matrix for the comprehensive evaluation index system of hydropower station power supply. The rows and columns of the matrix are all power supply status evaluation indicators. In the matrix, "1" indicates that the corresponding index in the row has an impact on the index in the column, and "0" indicates that there is no impact.
5. The method according to claim 3, characterized in that, In 2-2), the ANP network layer indicators are scored according to the indirect dominance principle to construct a judgment matrix. The specific method is as follows: Suppose that the control layer of the ANP network structure of the power plant auxiliary power system of a hydropower station has m indices B1, B2, B3, ..., Bm, and the network layer has N indices C1, C2, C3, ..., Cn. Taking a certain index Ci of the network layer as the criterion, the importance of all indices affecting Ci under the corresponding control layer Bm is compared pairwise. The nine-fold scaling method is used to construct the ANP judgment matrix for Ci under the control layer Bm. By analogy, the ANP judgment matrices of all control layers Bm and network layers Ci are constructed.
6. The method according to claim 5, characterized in that, In 2-3), construct the supermatrix. The specific steps are as follows: First, the eigenvectors of the normalized judgment matrix are calculated using the eigenvalue method, which are the weight vectors of the judgment matrix. After the consistency check passes, the normalized eigenvectors obtained from the judgment matrices of all indicators within the control layer Bm are combined to obtain the weight vector matrix. : ； in, The column vector in the text is a normalized feature vector calculated by using a certain indicator in the multi-dimensional indicator system Bm of hydropower plant power consumption as a secondary criterion and constructing a judgment matrix using the indicators in Bi that affect that indicator. Formula for the consistency test coefficient (CR number): ； The default value is CR < 0.1, which is sufficient to pass the consistency test; where n is the order of each judgment matrix constructed for the hydropower station's power supply system. This is the largest eigenvalue of the judgment matrix in this verification. Secondly, repeat the above operation for all metrics within all network layers C, and integrate all calculated weight vector matrices to obtain a supermatrix. : ； Among them, supermatrix Each column is a combination of weight vectors representing the influence of indicators on a certain indicator in the multi-dimensional indicator system network layer of all hydropower plant power consumption, where n is the total number of indicator elements in the network layer.
7. The method according to claim 6, characterized in that, In section 2-4), the specific steps for constructing the weighted supermatrix are as follows: First, construct the weighted matrix A. Specifically, calculate the normalized eigenvectors of the control layer's judgment matrix and arrange and integrate them to obtain the weighted matrix A. ； Among them, the first column of the weighting matrix A is the weight vector of all control layer indicators of the multi-dimensional indicator system for power plant power of all hydropower stations, when the first control layer indicator is the secondary criterion, and so on. Secondly, regarding the supermatrix in steps 2-3) Weighting is performed using a weighting matrix A, that is, matrix A and the hypermatrix. Block multiplication yields a weighted hypermatrix : ； Among them, the weighted supermatrix This reflects the direct influence relationship between multiple indicators of power plant power consumption in hydropower stations and the importance weight between groups, where n is the total number of all indicator elements in the network layer.
8. The method according to claim 7, characterized in that, In section 2-5), the specific steps for constructing the limiting hypermatrix and obtaining the global weights of ANP are as follows: when When the limit hypermatrix is obtained, Multiplying itself to infinite powers until the matrix result converges, the column vectors of the limiting hypermatrix at this point are the global weights of the hydropower plant power system state assessment index obtained using the ANP method.
9. The method according to any one of claims 1 to 8, characterized in that, In step S3, based on the established multi-dimensional indicator system, specifically, data is collected in real time through multiple channels such as SCADA system, smart sensors, and equipment inspection terminals. Since the collected hydropower plant power data has different dimensions and positive and negative polarities, it needs to be standardized and preprocessed. Positive performance indicators for hydropower plant power, such as insulation resistance and protection operation accuracy, are processed using the following formulas: ； Negative power output indicators for hydropower plants, such as voltage deviation, harmonic distortion rate, and temperature rise, are specifically processed using the following formulas: ； in, This represents the original value of the k-th indicator for the i-th sample of the hydropower station's power supply system. , This is the standardized value.
10. The method according to claim 1, characterized in that, In step S4, a correlation coefficient matrix is constructed using grey relational analysis, thereby calculating the grey relational coefficients; the specific steps are as follows: 4-1) Let the standard level be... There are a total of 5 safety levels, and a standard reference sequence for various indicators of power plant power consumption data of hydropower stations is established; ; Among them, the above matrix Each row represents a status level, and each column represents the standard values of each level of indicators for hydropower plant power consumption after standardization. Set the set of sequences to be analyzed for association with the reference sequences as comparison sequences. , that is, the real-time data of the corresponding indicators of the power plant power of the hydropower station collected, m comparison sequences, each sequence having n values; 4-2) Calculate the grey relational coefficient between the actual data and standard data of the hydropower station's power supply system; According to the grey system theory, a grey system is a system in which some information is known and some information is unknown, and it is used to solve the uncertainty caused by incomplete information. The status assessment of the power plant system is assumed to be a grey system, and the safety assessment of the power plant system can be carried out through grey relational analysis (GRA). ； in, Let be the grey coefficient between the k-th indicator and the k-th optimal indicator of the i-th evaluation object in the plant power status evaluation. It is the gray coordination coefficient; ; ; This represents the absolute difference between the reference sequence and the comparison sequence at each time point. The minimum value of the difference sequence. It is the maximum value of the difference sequence.