Thermal power generation system efficiency measuring method based on DBSCAN algorithm and DEA model

Abstract

The invention discloses a thermal power generation system efficiency measurement method based on a DBSCAN algorithm and a DEA model. The method comprises the following steps: 1, constructing an index system for evaluating the efficiency of a thermal power generation system; 2, performing homogeneity test on each secondary index based on a DBSCAN algorithm to obtain a homogeneous thermal power generation system; 3, calculating a primary index comprehensive cross efficiency value of the homogeneous thermal power generation system by using a cross efficiency DEA model; 4, complementing each first-level index missing value based on a KNN algorithm; and 5, calculating the efficiency of the thermal power generation system based on a weighted comprehensive scoring method. According to the method, an evaluation system with high efficiency, high precision and high robustness is realized, and a scientific tool is provided for efficiency optimization and green development of a thermal power generation system.

Description

Technical Field

[0001] The present invention relates to the technical field of measuring the efficiency of thermal power generation systems, and particularly to a method for measuring the efficiency of thermal power generation systems based on the DBSCAN algorithm and the DEA model. Background Art

[0002] As one of the core ways of global energy supply, thermal power generation can provide stable and reliable electric energy, meet the electricity demands of industry, commerce and residential life, and support the development of the economic society. However, there are problems such as resource waste, low energy utilization efficiency, and excessive pollutant emissions in the production process of thermal power generation systems, which directly affect their long-term stable operation and sustainable development. Therefore, it is crucial to measure the efficiency of thermal power generation systems, so as to optimize weak links, rationally allocate resources, improve energy utilization efficiency, and reduce environmental pollution.

[0003] Traditional efficiency evaluation methods mostly adopt single indicators or simple comprehensive evaluation models, which are difficult to comprehensively reflect the multi-dimensional efficiency characteristics of the system, especially insufficient in the comprehensive consideration of energy efficiency, operation efficiency and environmental efficiency. Moreover, traditional methods have weak processing capabilities for outliers and noise data, which are extremely likely to lead to distorted evaluation results. Data Envelopment Analysis (DEA), as a non-parametric efficiency evaluation method, can effectively handle multi-input and multi-output problems, but it is sensitive to outliers and noise data and cannot automatically identify the clustering structure in the data. The density clustering algorithm (DBSCAN) has strong noise processing capabilities and clustering recognition capabilities, but lacks an efficiency evaluation function. Summary of the Invention

[0004] In order to solve the above-mentioned deficiencies and defects, the present invention proposes a method for measuring the efficiency of thermal power generation systems based on the DBSCAN algorithm and the DEA model, in order to provide an efficiency measurement tool that takes into account both accuracy and interpretability for thermal power generation systems, can more accurately evaluate the comprehensive efficiency of thermal power generation systems, thereby optimizing weak links, improving energy utilization efficiency, providing a scientific basis for system optimization and decision-making, guiding the direction of technological improvement, and contributing to green, low-carbon and sustainable development.

[0005] To achieve the above invention objective, the present invention adopts the following technical solutions:

[0006] The method for measuring the efficiency of a thermal power generation system based on the DBSCAN algorithm and the DEA model of the present invention is characterized by including the following steps:

[0007] S1. Construct an index system for evaluating the efficiency of a thermal power generation system, including: a first-level index and several second-level indexes, and let any j-th first-level index be denoted as A j ; the ones belonging to the j-th first-level index A jThe secondary index set is denoted as , which represents the j th secondary index of the j-th primary index A; represents the number of secondary indices of the j-th primary index A j ;

[0008] S2. Conduct homogeneity tests on each secondary index based on the DBSCAN algorithm to obtain homogeneous thermal power generation systems;

[0009] S3. Use the cross-efficiency DEA model to calculate the comprehensive cross-efficiency values of the homogeneous thermal power generation systems under each primary index;

[0010] S4. Based on the KNN algorithm, complete the missing values of the comprehensive cross-efficiency of the thermal power generation systems under each primary index;

[0011] S5. Calculate the efficiency of the k-th thermal power generation system using Equation (7) ;

[0012] (7)

[0013] In Equation (7), represents the efficiency weight of the j-th primary index A j , represents the comprehensive cross-efficiency value after completion of the k-th thermal power generation system under the j-th primary index A j ;

[0014] The characteristics of the method for measuring the efficiency of a thermal power generation system based on the DBSCAN algorithm and the DEA model described in the present invention also lie in that the step S2 includes:

[0015] S2.1. Obtain the data of m thermal power generation systems to be evaluated corresponding to each secondary index under the j-th primary index A j to construct the original data matrix of the j-th dimension , where represents the data of the k-th thermal power generation system of the j th secondary index under the j-th primary index A;

[0016] S2.2. Let the k-th row vector in be represented as the k-th data point in the vector space, so as to obtain a data point set composed of data points;

[0017] S2.3. Set the parameters of the DBSCAN algorithm, including: the neighborhood radius Eps and the minimum number of points ;

[0018] S2.4. Calculate the Euclidean distance between in and the q-th data point using Equation (1), and then calculate the neighborhood of using Equation (2); ;

[0019] (1)

[0020] (2)

[0021] In Equation (1), represents the q-th thermal power generation system data of the j -th secondary index under the j-th primary index A; ;

[0022] S2.5. If , mark as a core point, and construct a clustering cluster with as the clustering center and its neighborhood as the elements of the cluster. If there is a core point in the neighborhood of , merge the cluster where the core point is located with the cluster where is located; otherwise, mark as a noise point and delete it. Here, represents the number of points counted in ;

[0023] S2.6. Traverse all the data points in according to the process of S2.4 - S2.5, so as to obtain all the clustering clusters corresponding to the core points and the data matrix after removing the noise points, , and regard the thermal power generation systems corresponding to the data points within the clustering cluster as homogeneous thermal power generation systems; regard the thermal power generation systems corresponding to the noise points as heterogeneous thermal power generation systems. Here, represents the j -th homogeneous thermal power generation system data of the -th secondary index under the j-th primary index A; ; represents the total number of homogeneous thermal power generation systems.

[0024] Furthermore, the step S3 includes:

[0025] S3.1. Divide each secondary indicator under the j-th first-level indicator A j into input indicators and output indicators, and according to the expected and non-expected attributes of the input and output indicators, the -th input indicator with expected attributes is denoted as , the -th output indicator with expected attributes is denoted as , the -th input indicator with non-expected attributes is denoted as , the -th output indicator with non-expected attributes is denoted as , thus obtaining the input-output data matrix for the j-th dimension; where ; among them, represents the number of input indicators with expected attributes, represents the number of input indicators with non-expected attributes, represents the number of output indicators with expected attributes, represents the number of output indicators with non-expected attributes;

[0026] S3.2. Use Equation (3) to construct the DEA self-evaluation efficiency model for the d-th homogeneous thermal power generation system;

[0027] (3)

[0028] In Equation (3), represents the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A j , , respectively represent the output value of the j -th expected attribute and the output value of the -th non-expected attribute of the d-th homogeneous thermal power generation system under the j-th first-level indicator A , , respectively represent the input value of the j -th expected attribute and the input value of the -th non-expected attribute of the d-th homogeneous thermal power generation system under the j-th first-level indicator A , , respectively represent the j -th expected attribute output indicator and the -th non-expected attribute output indicator when measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A The weight coefficient of a non - desired attribute output index and respectively represent the weight coefficients of the j th desired attribute input index and the th non - desired attribute input index when measuring the self - evaluation efficiency value of the d - th homogeneous thermal power generation system under the j - th first - level indicator A ;

[0029] S3.3. Use the Charnes - Cooper transformation to linearize Equation (3) to obtain the linearized multiplier - form model, and use the linear programming solution method to solve it. Correspondingly, obtain the optimal weight coefficients of the j th desired attribute output index and the optimal weight coefficients of the th non - desired attribute output index when measuring the self - evaluation efficiency value of the d - th homogeneous thermal power generation system under the j - th first - level indicator A , the optimal weight coefficients of the th desired attribute input index and the optimal weight coefficients of the th non - desired attribute input index when measuring the self - evaluation efficiency value of the d - th homogeneous thermal power generation system under the j - th first - level indicator A j ;

[0030]

[0031] S3.4. Use Equation (4) to calculate the cross - efficiency value j of the th homogeneous thermal power generation system and the d - th homogeneous thermal power generation system under the j - th first - level indicator A ;

[0031] (4)

[0032] In Equation (4), and respectively represent the optimal weight coefficients of the j th desired attribute output index and the th non - desired attribute output index when measuring the self - evaluation efficiency value of the th homogeneous thermal power generation system under the j - th first - level indicator A , and respectively represent the optimal weight coefficients of the th desired attribute output index and the th non - desired attribute output index when measuring the self - evaluation efficiency value of the j th homogeneous thermal power generation system under the j - th first - level indicator A ; Output indicators of expected attributes and the output indicators of non-expected attributes optimal weight coefficients of ;

[0033] S3.5. Calculate the comprehensive cross-efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A using Equation (5) j ;

[0034] (5)

[0035] S3.6. Calculate the comprehensive cross-efficiency values of all homogeneous thermal power generation systems under the j-th first-level indicator A according to the process of S3.2 - S3.5, so as to obtain j the comprehensive cross-efficiency values of all homogeneous thermal power generation systems under the j-th first-level indicator.

[0036] Furthermore, the step S3.3 includes:

[0037] S3.3.1. Obtain the linearized multiplier form model using Equation (8);

[0038] (8)

[0039] In Equation (8), t represents the proportionality factor, and ; represents the j -th transformed weight coefficient of the output indicators of expected attributes when measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A ; represents the -th transformed weight coefficient of the output indicators of non-expected attributes when measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A j ; represents the -th transformed weight coefficient of the input indicators of expected attributes when measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A ; j represents the -th transformed weight coefficient of the input indicators of non-expected attributes when measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A ; j ; represents the -th transformed weight coefficient of the input indicators of non-expected attributes when measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A

[0040] ​​S3.3.2. Solve Equation (8) to obtain the j-th first-level indicator A j When measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system, the expected attribute output index transformed optimal weight coefficient , the j-th first-level indicator A j When measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system, the undesired attribute output index transformed optimal weight coefficient , the j-th first-level indicator A j When measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system, the expected attribute input index transformed optimal weight coefficient , the j-th first-level indicator A j When measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system, the undesired attribute input index transformed optimal weight coefficient ;

[0041] S3.3.3. Use Equation (9) to obtain ;

[0042] (9).

[0043] Furthermore, the step S4 includes:

[0044] S4.1. Construct a comprehensive cross-efficiency matrix of dimension from the comprehensive cross-efficiency values of all homogeneous thermal power generation systems under the first-level indicators and the missing comprehensive cross-efficiency values of heterogeneous thermal power generation systems, where , represents the comprehensive cross-efficiency value of the k-th thermal power generation system under the j-th first-level indicator A j ; when , let , otherwise is a null value; represents the number of first-level indicators;

[0045] S4.2. Denote the column vector with null values in the -th column of E as the -th data point in the vector space , and denote the column vector without null values in the -th column of E as the a data point , calculate the distance between two data points using Equation (6) ;

[0046] (6)

[0047] In Equation (6), represents the distance weight of the j-th first-level index A j ;

[0048] S4.3. Set the parameter K, use the KNN algorithm to calculate the K data points without null values that are closest to the data point with null values, and use the mean value of the K data points without null values to fill in the null values of the data point, so as to obtain the completed matrix ; where represents the comprehensive cross-efficiency value of the k-th thermal power generation system after completion under the j-th first-level index A j .

[0049] An electronic device of the present invention includes a memory and a processor, characterized in that the memory is used to store a program that supports the processor to execute the method for measuring the efficiency of the thermal power generation system, and the processor is configured to execute the program stored in the memory.

[0050] A computer-readable storage medium of the present invention stores a computer program, characterized in that the computer program executes the steps of the method for measuring the efficiency of the thermal power generation system when run by a processor.

[0051] Compared with the existing technology, the beneficial effects of the present invention are as follows:

[0052] 1. By combining the advantages of the DBSCAN algorithm and the DEA model, the present invention proposes an efficient and accurate method for measuring the efficiency of the thermal power generation system. By comprehensively evaluating the energy efficiency, operation efficiency and environmental efficiency, the comprehensive efficiency of the thermal power generation system is calculated, overcoming the deficiencies of traditional methods in multi-dimensional efficiency evaluation, being able to further identify the reasons for the low system efficiency, thereby optimizing the weak links pertinently, improving the energy utilization efficiency, reducing environmental pollution, and making plans for technological improvement and upgrading, thus promoting the development of the thermal power generation field.

[0053] 2. The present invention introduces a cross-DEA model in the evaluation of the efficiency of thermal power generation systems, reducing the possible weight manipulation problems in traditional DEA, providing more comprehensive analysis results, making the evaluation results more objective, fair, and discriminatory. And a homogeneity test is carried out before efficiency evaluation. The DBSCAN algorithm has strong noise processing ability, avoiding the efficiency evaluation deviation caused by heterogeneity, and optimizing the effectiveness and robustness of the DEA model. This method is scientific and reasonable, and the evaluation conclusion is more in line with the actual situation, and can more comprehensively, accurately, and scientifically reflect the performance of each index of the thermal power generation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 is a flowchart of the method for measuring the efficiency of the thermal power generation system of the present invention;

[0055] Figure 2 is a flowchart of the homogeneity test for the secondary index set of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0056] In this example, a method for measuring the efficiency of a thermal power generation system based on the DBSCAN algorithm and the DEA model is as Figure 1 shown. First, an index system for evaluating the efficiency of the thermal power generation system is constructed; secondly, a homogeneity test is carried out on each secondary index based on DBSCAN to obtain a homogeneous thermal power generation system; then, the cross-efficiency DEA model is used to calculate the comprehensive cross-efficiency value of the primary index of the homogeneous thermal power generation system; then, the missing values of each primary index are complemented based on the KNN algorithm; finally, the efficiency of the thermal power generation system is calculated based on the weighted comprehensive scoring method. Specifically, the steps of the method for measuring the efficiency of the thermal power generation system based on the DBSCAN algorithm and the DEA model are as follows:

[0057] S1. Construct an index system for evaluating the efficiency of the thermal power generation system, including: a number of primary indexes and a number of secondary indexes. The primary indexes include: energy efficiency A1, economic efficiency A2, and environmental efficiency A3; let any j-th primary index be denoted as A j , where represents the number of primary indexes, ; the secondary index set belonging to the j-th primary index A j is denoted as , represents the j -th secondary index of the j-th primary index A represents the number of secondary indexes of the j-th primary index A j ; the index system for measuring the efficiency of the thermal power generation system is shown in Table 1.

[0058] Table Index System for Measuring the Efficiency of Thermal Power Generation Systems

[0059]

[0060] S2. As Figure 2 shown, perform a homogeneity test on each secondary indicator based on the DBSCAN algorithm to obtain homogeneous thermal power generation systems;

[0061] S2.1. According to the scenarios and requirements of the DBSCAN algorithm, it is necessary to obtain the data of m thermal power generation systems to be evaluated corresponding to each secondary indicator under the j-th primary indicator A j to construct the original data matrix of dimension j, where represents the data of the k-th thermal power generation system for the j-th primary indicator A j under the -th secondary indicator ;

[0062] S2.2. Let the k-th row vector in be represented as the k-th data point in the vector space, thus obtaining the data point set consisting of data points.

[0063] S2.3. Set the parameters of the DBSCAN algorithm, including: neighborhood radius Eps and minimum number of points ;

[0064] S2.4. Calculate the Euclidean distance between and the q-th data point in using Equation (1), and then calculate the neighborhood of using Equation (2);

[0065] (1)

[0066] (2)

[0067] In Equation (1), represents the data of the q-th thermal power generation system for the j -th secondary indicator under the j-th primary indicator A ;

[0068] S2.5. If , then Mark as the core points and construct a as the clustering center, and use its neighborhood as the elements of a cluster. If there is a core point in the neighborhood of , then merge the cluster where the core point is located with the -located cluster; otherwise, mark as a noise point and delete it; where represents the statistical number of points in

[0069] S2.6. Traverse all the data points in according to the process of S2.4 - S2.5, so as to obtain all the clusters corresponding to the core points and the data matrix after removing the noise points, . And regard the thermal power generation systems corresponding to the data points within the cluster as homogeneous thermal power generation systems; regard the thermal power generation systems corresponding to the noise points as heterogeneous thermal power generation systems; where represents the jth first-level indicator A j under the th second-level indicator of the th homogeneous thermal power generation system data; represents the total number of homogeneous thermal power generation systems;

[0070] S3. Calculate the comprehensive cross-efficiency value of the first-level indicators of the homogeneous thermal power generation system using the cross-efficiency DEA model;

[0071] S3.1. According to the requirements and usage scenarios of the DEA method, divide the secondary indicators under the jth first-level indicator A j into input indicators and output indicators, and according to the expected and non-expected attributes of the input indicators and output indicators, mark the th expected-attribute input indicator of as th expected-attribute output indicator of as th non-expected-attribute input indicator of as th non-expected-attribute output indicator of , so as to obtain the input-output data matrix of the jth dimension ; where represents the number of expected-attribute input indicators, represents the number of non-expected-attribute input indicators, represents the number of expected attribute output indicators, represents the number of unexpected attribute output indicators.

[0072] S3.2. Construct the DEA self-evaluation efficiency model of the d-th homogeneous thermal power generation system using Equation (3);

[0073] (3)

[0074] In Equation (3), represents the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A j , , respectively represent the output value of the j -th expected attribute and the output value of the -th unexpected attribute of the d-th homogeneous thermal power generation system under the j-th first-level indicator A , , respectively represent the input value of the j -th expected attribute and the input value of the -th unexpected attribute of the d-th homogeneous thermal power generation system under the j-th first-level indicator A , , respectively represent the weight coefficients of the j -th expected attribute output indicator and the -th unexpected attribute output indicator when measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A , , respectively represent the weight coefficients of the j -th expected attribute input indicator and the -th unexpected attribute input indicator when measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A .

[0075] S3.3. Use the Charnes-Cooper transformation to linearize Equation (3) to obtain the linearized multiplier form model, and use the linear programming solution method to solve it, and accordingly obtain the optimal weight coefficients j of the -th expected attribute output indicator and the optimal weight coefficients of the -th unexpected attribute output indicator when measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A , and the j-th first-level indicator A jWhen measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system, the optimal weight coefficient of the expected attribute input index and the optimal weight coefficient of the undesired attribute input index ; The specific process of linear transformation and solution is as follows:

[0076] S3.3.1. Obtain the linearized multiplier form model using Equation (4);

[0077] (4)

[0078] In Equation (4), t represents the scaling factor, and ; represents the j transformed weight coefficient of the expected attribute output index when measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A ; represents the j transformed weight coefficient of the undesired attribute output index when measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A ; represents the j transformed weight coefficient of the expected attribute input index when measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A ; represents the j transformed weight coefficient of the undesired attribute input index when measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A ;

[0079] S3.3.2. Use the linear programming solution method to solve Equation (4) to obtain the j transformed optimal weight coefficient of the expected attribute output index when measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A , the transformed optimal weight coefficient of the j undesired attribute output index when measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A , and the transformed optimal weight coefficient of the undesired attribute input index when measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A jWhen measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system, the optimal weight coefficient after transformation of the j-th expected attribute input index , and the optimal weight coefficient after transformation of the j k-th non-expected attribute input index when measuring the self-evaluation efficiency value of the d-th homogeneous thermal power generation system . optimal weight coefficient after transformation .

[0080] S3.3.3. Obtain using Equation (5);

[0081] (5)

[0082] S3.4. Calculate the cross-efficiency value between the j-th homogeneous thermal power generation system and the d-th homogeneous thermal power generation system using Equation (6); j at the j-th first-level indicator A ;

[0083] (6)

[0084] In Equation (6), , respectively represent the optimal weight coefficients of the j m-th expected attribute output index and the n-th non-expected attribute output index when measuring the self-evaluation efficiency value of the i-th homogeneous thermal power generation system under the j-th first-level indicator A ; , respectively represent the optimal weight coefficients of the m'-th expected attribute output index and the n'-th non-expected attribute output index when measuring the self-evaluation efficiency value of the j i'-th homogeneous thermal power generation system under the j-th first-level indicator A ; optimal weight coefficient of the m-th expected attribute output index and the n-th non-expected attribute output index . .

[0085] S3.5. Calculate the comprehensive cross-efficiency value of the d-th homogeneous thermal power generation system under the j-th first-level indicator A using Equation (7); j at ;

[0086] (7)

[0087] where , if indicates that the d-th homogeneous thermal power generation system is efficient under the j-th first-level indicator A j . If is less than 1, it means that the d-th homogeneous thermal power generation system is ineffective under the j-th first-level indicator A j .

[0088] S3.6. Calculate the comprehensive cross-efficiency values of all homogeneous thermal power generation systems under the j-th first-level indicator A j according to the process in S3.2 - S3.5, so as to obtain the comprehensive cross-efficiency values of all homogeneous thermal power generation systems under each first-level indicator.

[0089] S4. Complement the missing values of each first-level indicator based on the KNN algorithm;

[0090] S4.1. According to the scenarios and requirements of the KNN algorithm, construct a comprehensive cross-efficiency matrix of dimension using the comprehensive cross-efficiency values of all homogeneous thermal power generation systems under each first-level indicator and the missing comprehensive cross-efficiency values of heterogeneous thermal power generation systems, where represents the comprehensive cross-efficiency value of the k-th thermal power generation system under the j-th first-level indicator A . When let j , otherwise is a null value.

[0091] S4.2. Denote the column vector with null values in the -th column of E as the -th data point in the vector space . Denote the column vector without null values in the -th column of E as the -th data point in the vector space . Calculate the distance between the two data points using Equation (8);

[0092] (8)

[0093] In Equation (8), represents the distance weight of the j-th first-level indicator A j . If is a null value, then , otherwise .

[0094] ​​S4.3. Set parameter K, use the KNN algorithm to calculate the K data points without null values that are closest to the data points with null values, and use the mean of the K data points without null values to fill in the null values of the data points, so as to obtain the completed matrix. ; among them, represents the comprehensive cross - efficiency value after completion of the k - th thermal power generation system under the j - th first - level indicator A j .

[0095] S5. Calculate the efficiency of the k - th thermal power generation system using Equation (9) ;

[0096] (9)

[0097] In Equation (9), represents the efficiency weight of the j - th first - level indicator A j .

[0098] In this embodiment, an electronic device includes a memory and a processor. The memory is used to store a program that supports the processor to execute the above - mentioned method, and the processor is configured to execute the program stored in the memory.

[0099] In this embodiment, a computer - readable storage medium stores a computer program, and when the computer program is run by a processor, it executes the steps of the above - mentioned method.

[0100] In summary, this method effectively processes noise data and identifies the data clustering structure through the DBSCAN algorithm, comprehensively evaluates the efficiency of a multi - input multi - output system in combination with the DEA model, comprehensively calculates the energy efficiency, operation efficiency, and environmental efficiency, and finally obtains the comprehensive efficiency of the thermal power generation system. This method overcomes the deficiencies of traditional efficiency measurement methods in outlier processing, clustering identification, and multi - dimensional efficiency evaluation, significantly improves the accuracy and reliability of efficiency measurement, provides a scientific basis for the optimal operation, resource allocation, technical improvement and upgrading, and sustainable development of thermal power generation systems, and has important practical application value.

Claims

1. A method for measuring the efficiency of a thermal power generation system based on the DBSCAN algorithm and the DEA model, characterized in that: The following steps are involved: S1. Construct an index system for evaluating the efficiency of thermal power generation systems, including: first-level indicators and several second-level indicators, let any j-th first-level indicator be A j ; will belong to the jth first-level indicator A j The secondary index set is recorded as , Represents the jth first-level index A j No. Secondary indicators; Represents the jth first-level index A j The number of secondary indicators; S2, based on the DBSCAN algorithm, the homogeneity test of each secondary index is carried out to obtain a homogeneous thermal power generation system; S3. Use the cross-efficiency DEA model to calculate the comprehensive cross-efficiency value of the homogeneous thermal power generation system under each primary index; S4. Complete the missing values ​​of the comprehensive cross efficiency of the thermal power generation system under each primary indicator based on the KNN algorithm; S5. Calculate the efficiency of the kth thermal power generation system using formula (7): ; (7) In formula (7), Represents the jth first-level index A j The efficiency weight of Represents the jth first-level index A j The comprehensive cross efficiency value of the kth thermal power generation system after completion.

2. According to claim 1, a method for measuring the efficiency of a thermal power generation system based on a DBSCAN algorithm and a DEA model is characterized in that: The step S2 comprises: S2.

1. Obtain the jth primary index A j The m thermal power generation system data to be evaluated corresponding to each secondary indicator under , thus constructing the jth dimension The original data matrix ,in, Represents the jth first-level index A j Next Secondary indicators The kth thermal power generation system data; S2.

2. Order The kth row vector in Represented as the kth data point in vector space , thus obtaining The data point set consists of ; S2.

3. Set the parameters of the DBSCAN algorithm, including: neighborhood radius Eps and minimum number of points ; S2.4, Calculate using formula (1) middle With the qth data point The Euclidean distance between , and then use formula (2) to calculate Neighborhood ; (1) (2) In formula (1), Represents the jth first-level index A j Next Secondary indicators Data of the qth thermal power generation system; S2.5, if , then Mark as core points and construct is the cluster center, and its neighborhood is a cluster of elements of a cluster, if Neighborhood If there is a core point in the cluster, the cluster where the core point is located is The clusters where are located are merged; otherwise, Mark as noise points and delete them; among them, Representation Statistics the number of midpoints; S2.

6. Follow the process of S2.4-S2.5 to Traverse all the data points in the matrix to obtain the clusters corresponding to all the core points and the data matrix after removing the noise points , , and the thermal power generation system corresponding to the data points in the cluster is regarded as a homogeneous thermal power generation system; the thermal power generation system corresponding to the noise point is regarded as a heterogeneous thermal power generation system; where, Represents the jth first-level index A j Next Secondary indicators No. Data of homogeneous thermal power generation systems; Represents the total number of homogeneous thermal power generation systems.

3. The method for measuring the efficiency of a thermal power generation system based on the DBSCAN algorithm and the DEA model according to claim 2 is characterized in that: The step S3 comprises: S3.

1. The jth first-level index A j The secondary indicators under the above are divided into input indicators and output indicators, and according to the expected and undesirable attributes of input indicators and output indicators, No. The input index of the desired attribute is , No. The output index of the desired attribute is , No. The input index of the undesirable attribute is , No. The output index of the undesirable attribute is , so that the j-th dimension is Input-output data matrix ;in, Indicates the number of expected attribute input indicators, represents the number of non-desirable attribute input indicators, Indicates the number of expected attribute output indicators, Indicates the number of undesirable attribute output indicators; S3.2, using formula (3) to construct the DEA self-evaluation efficiency model of the dth homogeneous thermal power generation system; (3) In formula (3), Represents the jth first-level index A j The self-evaluated efficiency value of the dth homogeneous thermal power generation system is: , Respectively represent the jth first-level index A j The dth homogeneous thermal power generation system The output value of the desired attribute, The output value of the undesirable attribute, , Respectively represent the jth first-level index A j The dth homogeneous thermal power generation system The input value of the desired attribute, The input value of the undesirable attribute, , Respectively represent the jth first-level index A j When measuring the self-evaluated efficiency value of the dth homogeneous thermal power generation system, The expected attribute output indicator, The weight coefficient of the output index of the non-desirable attribute, , Respectively represent the jth first-level index A j When measuring the self-evaluated efficiency value of the dth homogeneous thermal power generation system, The expected attribute input index, The weight coefficient of the non-desirable attribute input index; S3.3, use Charnes-Cooper transformation to linearize equation (3), obtain the linearized multiplier form model, and use the linear programming solution method to solve it, and get the jth first-level index A accordingly j When measuring the self-evaluated efficiency value of the dth homogeneous thermal power generation system, Desired attribute output index The optimal weight coefficient , No. Undesirable attribute output index The optimal weight coefficient , the jth first-level index A j When measuring the self-evaluated efficiency value of the dth homogeneous thermal power generation system, Desired attribute input index The optimal weight coefficient , No. Undesirable attribute input index The optimal weight coefficient ; S3.

4. Calculate the jth primary index A using formula (4): j Next The cross efficiency value of the homogeneous thermal power generation system and the dth homogeneous thermal power generation system ; (4) In formula (4), , Respectively represent the jth first-level index A j Next measurement The self-evaluated efficiency value of a homogeneous thermal power generation system is Output indicators of desired attributes , No. Output indicators of undesirable attributes The optimal weight coefficient of , The jth first-level index A j Next measurement The self-evaluated efficiency value of a homogeneous thermal power generation system is Output indicators of desired attributes , No. Output indicators of undesirable attributes The optimal weight coefficient of ; S3.

5. Calculate the jth primary index A using formula (5): j The comprehensive cross efficiency value of the dth homogeneous thermal power generation system is ; (5) S3.

6. Calculate the jth primary index A according to the process of S3.2-S3.5 j The comprehensive cross efficiency value of all homogeneous thermal power generation systems under The comprehensive cross efficiency value of all homogeneous thermal power generation systems under the first-level indicators.

4. The method for measuring the efficiency of a thermal power generation system based on the DBSCAN algorithm and the DEA model according to claim 3 is characterized in that: The step S3.3 comprises: S3.3.1, use formula (8) to obtain the linearized multiplier form model; (8) In formula (8), t represents the proportional factor, and ; Represents the jth first-level index A j The self-evaluated efficiency value of the dth homogeneous thermal power generation system is measured at Desired attribute output index The transformed weight coefficient; Represents the jth first-level index A j When measuring the self-evaluated efficiency value of the dth homogeneous thermal power generation system, Undesirable attribute output index The transformed weight coefficient; Represents the jth first-level index A j When measuring the self-evaluated efficiency value of the dth homogeneous thermal power generation system, Desired attribute input index The transformed weight coefficient; Represents the jth first-level index A j When measuring the self-evaluated efficiency value of the dth homogeneous thermal power generation system, Undesirable attribute input index The transformed weight coefficient; S3.3.

2. Solve equation (8) to obtain the jth primary index A j When measuring the self-evaluated efficiency value of the dth homogeneous thermal power generation system, Desired attribute output index The optimal weight coefficient after transformation , the jth first-level index A j When measuring the self-evaluated efficiency value of the dth homogeneous thermal power generation system, Undesirable attribute output index The optimal weight coefficient after transformation , the jth first-level index A j When measuring the self-evaluated efficiency value of the dth homogeneous thermal power generation system, Desired attribute input index The optimal weight coefficient after transformation , the jth first-level index A j When measuring the self-evaluated efficiency value of the dth homogeneous thermal power generation system, Undesirable attribute input index The optimal weight coefficient after transformation ; S3.3.3, using formula (9) to obtain ; (9)。 5. The method for measuring the efficiency of a thermal power generation system based on the DBSCAN algorithm and the DEA model according to claim 4 is characterized in that: The step S4 comprises: S4.

1. The construction dimension of the comprehensive cross efficiency value of all homogeneous thermal power generation systems under the first-level index and the missing value of the comprehensive cross efficiency of heterogeneous thermal power generation systems is The comprehensive cross efficiency matrix ,in, Represents the jth first-level index A j The comprehensive cross efficiency value of the kth thermal power generation system under season ,otherwise, is a null value; Indicates the number of first-level indicators; S4.2, let E Column vector containing null values ​​for the columns In the vector space, it is denoted as Data points , let the first A column vector containing columns that do not have null values. In the vector space, it is denoted as Data points , use formula (6) to calculate the distance between two data points ; (6) In formula (6), Represents the jth first-level index A j The distance weight of S4.3, set the parameter K, use the KNN algorithm to calculate the K non-null data points closest to the null data points, and use the mean of the K non-null data points to fill in the null data points, so as to obtain the completed matrix ;in, Represents the jth first-level index A j The comprehensive cross efficiency value of the kth thermal power generation system after completion.

6. An electronic device, comprising a memory and a processor, characterized in that: The memory is used to store a program that supports the processor to execute the method for measuring the efficiency of a thermal power generation system as described in any one of claims 1 to 5, and the processor is configured to execute the program stored in the memory.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for measuring efficiency of a thermal power generation system according to any one of claims 1 to 5 are executed.

Images