Electric power data fingerprint evaluation system based on neural network model

Abstract

The invention discloses an electric power data fingerprint performance evaluation system based on a neural network model, relates to an electric power data fingerprint performance evaluation device, and aims to solve the problem that an existing evaluation method cannot objectively reflect information of electric power data fingerprint performance, so that an electric power data fingerprint performance evaluation result is inaccurate. A data selection module is used for constructing an alternative evaluation index set, quantifying indexes and determining classical domain matter elements; the data processing module is used for obtaining related evaluation data and determining matter elements to be evaluated; the weight determination module is used for determining an index objective weight and an index subjective weight; the weight optimization module is used for carrying out weight coupling optimization on the index objective weight and the index subjective weight to obtain a weight optimal solution; and the result output module uses the classical domain matter element and the matter element to be evaluated and combines the weight optimal solution to calculate the grade rating correlation degree, and evaluates the fingerprint performance of the power data. The method has the beneficial effect that the accuracy of an electric power data fingerprint evaluation result is improved.

Description

Technical Field

[0001] The invention relates to a power data fingerprint evaluation device. Background Art

[0002] The fingerprint of power data is one of the important technologies for the application of smart substations. The application of the fingerprint of power data indicates that the power industry has made a major breakthrough in ensuring the security of equipment operation data and preventing errors and tampering. The power industry is related to the national economy and people's livelihood. Ensuring the security of equipment operation data and resisting malicious tampering from the outside world are the key to ensuring power supply and benefiting people's livelihood. After the fingerprint of power data is applied, the smart substation has the two major advantages of "safety" and "efficiency". The fingerprint of power data has the following conventional indicators: capacity, robustness, imperceptibility, reliability, and anti-exploitation. In real scenarios, it is also necessary to consider the impact of the embedded position on the data itself, the encryption and decryption calculation efficiency of the fingerprint of power data, the versatility of the fingerprint scheme of power data, and expert opinions. For the above multiple indicators, the existing single evaluation methods such as the analytic hierarchy process and the entropy weight method cannot accurately and objectively complete the evaluation task, nor can they objectively reflect the information of the fingerprint of power data itself. Summary of the invention

[0003] The purpose of the present invention is to solve the problem that the existing evaluation method cannot objectively reflect the information of power data fingerprint itself, resulting in inaccurate power data fingerprint evaluation results, and proposes a power data fingerprint evaluation system based on a neural network model.

[0004] The invention discloses a power data fingerprint evaluation system based on a neural network model, wherein the power data fingerprint evaluation system comprises a data selection module, a data processing module, a weight determination module, a weight optimization module and a result output module;

[0005] The data selection module is used to construct an alternative evaluation indicator set based on the daily use scenarios of the power database and the evaluation indicators of the digital fingerprint, and to quantify the indicators and determine the classical domain matter-element;

[0006] The data processing module is used to construct a candidate evaluation index set according to the data selection module to obtain relevant evaluation data and determine the matter-element to be evaluated;

[0007] A weight determination module is used to determine the objective weight of the indicator and the subjective weight of the indicator;

[0008] The weight optimization module uses the optimal weighting method and the Lagrange multiplier method to couple and optimize the objective weights of the indicators determined by the weight determination module and the subjective weights of the indicators to obtain the optimal weight solution;

[0009] The result output module uses the classical domain matter-elements determined by the data selection module and the matter-elements to be evaluated determined by the data processing module, and combines the weighted optimal solution obtained by the weight optimization module to calculate the grade rating correlation and evaluate the fingerprint of the power data.

[0010] Furthermore, the evaluation indicators in the data selection module include capacity evaluation, robustness evaluation, imperceptibility evaluation, reliability evaluation and anti-exploitation evaluation.

[0011] Further, the capacity evaluation is: the maximum length of the fingerprint that can be embedded;

[0012] The robustness evaluation includes deletion attack, addition attack, modification attack, protocol attack and synchronization attack;

[0013] The imperceptibility evaluation includes modifying the number of rows, variance, mean, and the absolute value of the difference between the embedded column and the original column over the value of the original column;

[0014] The reliability evaluation includes missed detection rate, false detection rate, traceability and provability;

[0015] The anti-exploitation assessment includes communication confidentiality and encryption algorithm confidentiality.

[0016] Furthermore, the data processing module includes an evaluation module and a matter-element determination module;

[0017] The evaluation module is used to construct an alternative evaluation indicator set based on the data selection module to obtain relevant evaluation data.

[0018] The matter-element determination module is used to determine the matter-element to be evaluated according to the indicator quantification method.

[0019] Furthermore, the indicator quantification method of the matter-element determination module is:

[0020] The data selection module is used to construct an alternative set of evaluation indicators, and a single indicator evaluation system is established, that is, the relationship between the quantified measurement value and the evaluation interval is determined. The primitive theory in extenics is used to determine the classical domain object element. For extremely small, intermediate, and interval indicators, relevant positive methods are used to adjust them to extremely large indicators.

[0021] Further, the weight determination module includes an objective weight module and a subjective weight module;

[0022] The objective weight module is used to determine the objective weight of the indicator using the entropy weight method;

[0023] The subjective weight module is used to determine the subjective weight of indicators using the hierarchical analysis method.

[0024] Compared with the prior art, the present invention has the following beneficial effects:

[0025] The power data fingerprint evaluation system based on the neural network model described in the present invention combines the characteristics of power data use and the collusion phenomenon that exists in reality. It uses the framework structure of the neural network model and the basic principles of the neural network to propose traceability and anti-collusion indicators for power data fingerprints, making the evaluation system objective and reasonable. By establishing classical domain matter-element, the evaluation object can be more accurately quantitatively described; and the objective weight of the indicator is determined by the entropy weight method, the subjective weight of the indicator is determined by the hierarchical analysis method, and the optimal weighting method is used to couple and optimize the weights of the objective weight of the indicator and the subjective weight of the indicator, so that the evaluation results are more realistic and the accuracy of the power data fingerprint evaluation results is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is a structural block diagram of a power data fingerprint evaluation system based on a neural network model as described in a specific implementation method 1. DETAILED DESCRIPTION

[0027] Specific implementation method 1. Combination Figure 1 The present embodiment is described. The power data fingerprint evaluation system based on the neural network model described in the present embodiment is characterized in that the power data fingerprint evaluation system includes a data selection module 1, a data processing module 2, a weight determination module 3, a weight optimization module 4 and a result output module 5;

[0028] Data selection module 1 is used to construct an alternative evaluation index set based on the daily use scenarios of the power database and the evaluation index of the digital fingerprint, quantify the index, and determine the classical domain matter-element;

[0029] Data processing module 2, used to construct a candidate evaluation index set according to data selection module 1 to obtain relevant evaluation data and determine the matter-element to be evaluated;

[0030] The weight determination module 3 is used to determine the objective weight of the indicator and the subjective weight of the indicator;

[0031] The weight optimization module 4 uses the optimal weighting method and the Lagrange multiplier method to couple and optimize the objective weight of the indicator and the subjective weight of the indicator determined by the weight determination module 3 to obtain the optimal weight solution;

[0032] The result output module 5 uses the classical domain matter-element determined by the data selection module 1 and the matter-element to be evaluated determined by the data processing module 2, and combines the weighted optimal solution obtained by the weight optimization module 4 to calculate the grade rating correlation and evaluate the fingerprint of the power data.

[0033] The power data fingerprint evaluation system based on the neural network model described in this embodiment combines the characteristics of power data use and the collusion phenomenon that exists in reality. It uses the framework structure of the neural network model and the basic principles of the neural network to propose traceability and anti-collusion indicators for power data fingerprints, making the evaluation system objective and reasonable. By establishing the classical domain object element, the evaluation object can be more accurately quantitatively described; and the objective weight of the indicator is determined by the entropy weight method, the subjective weight of the indicator is determined by the hierarchical analysis method, and the optimal weighting method is used to couple the weight optimization of the objective weight of the indicator and the subjective weight of the indicator, so that the evaluation results are more realistic and the accuracy of the power data fingerprint evaluation results is improved.

[0034] Specific implementation method 2: This implementation method further limits the power data fingerprint evaluation system based on the neural network model described in specific implementation method 1. In this implementation method, the evaluation indicators in the data selection module 1 include capacity evaluation, robustness evaluation, imperceptibility evaluation, reliability evaluation and anti-exploitation evaluation.

[0035] In this implementation, an alternative set of evaluation indicators is constructed. The daily use scenarios of the power database include mean query of specific columns, interval query of specific columns, update of intermittent time, large amount of data, etc., so it is necessary to focus on the efficiency and imperceptibility of the embedded algorithm; in real scenarios, there are many collusion phenomena, so it is necessary to consider the anti-collusion property of data fingerprints; there are high requirements for the traceability capability of data fingerprints, so traceability indicators are proposed.

[0036] Conventional evaluation indicators of digital fingerprints include capacity, robustness, imperceptibility, reliability, and anti-exploitation; among them, capacity evaluation is the maximum length of the embedded fingerprint; robustness evaluation includes deletion attack, addition attack, modification attack, protocol attack, and synchronization attack; imperceptibility evaluation includes the number of modified rows, variance, mean, and the absolute value of the difference between the embedded result and the original column divided by the value of the original column; reliability evaluation includes missed detection rate, false detection rate, traceability, and provability; anti-exploitation evaluation includes communication confidentiality, encryption algorithm confidentiality, etc.

[0037] Specific implementation method 3, this implementation method is to further limit the power data fingerprint evaluation system based on the neural network model described in specific implementation method 2. In this implementation method, the capacity evaluation is: the maximum length of the fingerprint that can be embedded;

[0038] The robustness evaluation includes deletion attack, addition attack, modification attack, protocol attack and synchronization attack;

[0039] The imperceptibility evaluation includes modifying the number of rows, variance, mean, and the absolute value of the difference between the embedded column and the original column over the value of the original column;

[0040] The reliability evaluation includes missed detection rate, false detection rate, traceability and provability;

[0041] The anti-exploitation assessment includes communication confidentiality and encryption algorithm confidentiality.

[0042] In this embodiment, capacity assessment is to examine the number of effective bits of data fingerprint; robustness assessment includes but is not limited to deletion attack, that is, deleting a certain proportion of tuples and observing the traceability of data fingerprint; addition attack, that is, adding a certain proportion of tuples and observing the traceability of data fingerprint; modification attack, that is, modifying a certain proportion of tuples and observing the traceability of data fingerprint; protocol attack, that is, trying to interpret the data from the user's perspective, observing the anti-interpretation ability of the data fingerprint, and finally determining the robustness assessment result; imperceptibility assessment is to examine the number of modified rows, the variance and mean comparison before and after the modification, or use the ratio of the absolute value of the difference between the embedded column and the original column to the value of the original column as a comprehensive indicator; reliability assessment is to examine the missed detection rate and false detection rate of the data fingerprint, and determine the corresponding evaluation result based on the mathematical proof of the provability of the data fingerprint; anti-exploitation assessment is to examine the encryption capability of the data fingerprint encoding algorithm itself, and determine the anti-exploitation assessment result based on the encryption standard.

[0043] Specific implementation method 4: This implementation method further limits the power data fingerprint evaluation system based on the neural network model described in specific implementation method 1. In this implementation method, the data processing module 2 includes an evaluation module and a matter-element determination module;

[0044] An evaluation module, used to construct an alternative evaluation indicator set according to the data selection module 1 to obtain relevant evaluation data;

[0045] The matter-element determination module is used to determine the matter-element to be evaluated according to the indicator quantification method.

[0046] Specific implementation method 5: This implementation method further limits the power data fingerprint evaluation system based on the neural network model described in specific implementation method 4. In this implementation method, the index quantification method of the matter-element determination module is:

[0047] The data selection module 1 is used to construct an alternative set of evaluation indicators and establish a single indicator evaluation system, that is, to quantify the relationship between the measurement value and the evaluation interval, and to use the primitive theory in extenics to determine the classical domain object element; for extremely small, intermediate, and interval indicators, the relevant positive method is used to adjust them to extremely large indicators.

[0048] Specific implementation method 6. This implementation method further limits the power data fingerprint evaluation system based on the neural network model described in specific implementation method 1. In this implementation method, the weight determination module 3 includes an objective weight module and a subjective weight module;

[0049] The objective weight module is used to determine the objective weight of the indicator using the entropy weight method;

[0050] The subjective weight module is used to determine the subjective weight of indicators using the hierarchical analysis method.

[0051] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A power data fingerprint evaluation system based on a neural network model, characterized in that: The power data fingerprint evaluation system comprises a data selection module (1), a data processing module (2), a weight determination module (3), a weight optimization module (4) and a result output module (5); A data selection module (1) is used to construct an alternative evaluation index set based on the daily use scenarios of the power database and the evaluation index of the digital fingerprint, and to quantify the index and determine the classical domain matter-element; The data processing module (2) is used to construct a candidate evaluation index set according to the data selection module (1) to obtain relevant evaluation data and determine the matter-element to be evaluated; A weight determination module (3) is used to determine the objective weight of the indicator and the subjective weight of the indicator; The weight optimization module (4) uses the optimal weighting method and the Lagrange multiplier method to couple and optimize the objective weight of the indicator and the subjective weight of the indicator determined by the weight determination module (3) to obtain the optimal weight solution; The result output module (5) calculates the grade rating correlation and evaluates the power data fingerprint by using the classical domain matter-element determined by the data selection module (1) and the matter-element to be evaluated determined by the data processing module (2) and combining the weighted optimal solution obtained by the weight optimization module (4).

2. According to the neural network model-based power data fingerprint evaluation system of claim 1, it is characterized in that: The evaluation indicators in the data selection module (1) include capacity evaluation, robustness evaluation, imperceptibility evaluation, reliability evaluation and anti-exploitation evaluation.

3. According to claim 2, a power data fingerprint evaluation system based on a neural network model is characterized in that: The capacity assessment is: the maximum length of the fingerprint that can be embedded; The robustness evaluation includes deletion attack, addition attack, modification attack, protocol attack and synchronization attack; The imperceptibility evaluation includes modifying the number of rows, variance, mean, and the absolute value of the difference between the embedded column and the original column over the value of the original column; The reliability evaluation includes missed detection rate, false detection rate, traceability and provability; The anti-exploitation assessment includes communication confidentiality and encryption algorithm confidentiality.

4. The power data fingerprint evaluation system based on the neural network model according to claim 1 is characterized in that: The data processing module (2) includes an evaluation module and a matter-element determination module; The evaluation module is used to construct an alternative evaluation indicator set based on the data selection module (1) to obtain relevant evaluation data. The matter-element determination module is used to determine the matter-element to be evaluated according to the indicator quantification method.

5. The power data fingerprint evaluation system based on the neural network model according to claim 4 is characterized in that: The index quantification method of the matter-element determination module is: The data selection module (1) is used to construct a set of alternative evaluation indicators and establish a single indicator evaluation system, that is, to quantify the relationship between the measurement value and the evaluation interval, and to use the primitive theory in extenics to determine the classical domain object element; for the extremely small, intermediate, and interval indicators, the relevant positive conversion method is used to adjust them to extremely large indicators.

6. The power data fingerprint evaluation system based on the neural network model according to claim 1 is characterized in that: The weight determination module (3) includes an objective weight module and a subjective weight module; The objective weight module is used to determine the objective weight of the indicator using the entropy weight method; The subjective weight module is used to determine the subjective weight of indicators using the hierarchical analysis method.

Images