CN-121997372-A - Dynamic privacy protection method and system for electric power data based on isomorphic encryption

CN121997372ACN 121997372 ACN121997372 ACN 121997372ACN-121997372-A

Abstract

The invention provides a dynamic privacy protection method and a dynamic privacy protection system for electric power data based on homomorphic encryption, which relate to the technical field of data processing and comprise the steps of acquiring electric power data and homomorphic encryption, and establishing an initial noise reference value; the method comprises the steps of constructing a ciphertext domain state evaluation model to analyze dynamic association characteristics to generate a hierarchical privacy protection strategy, constructing a noise sensitivity curve to obtain a noise threshold reference value, and determining noise attenuation parameters to optimize ciphertext operation. The invention solves the problem that privacy protection and data availability are difficult to balance in the traditional method, and realizes high-efficiency analysis and safety protection of the power data.

Inventors

WANG SINING
ZHANG LUFENG
WANG YAQIAN
ZHAO FENG
WEI ZHIFENG
Xia Baobing
GAO BINGQIANG
LIU HAIYANG
HAN YUXIN
ZHAO ZHENXIA
ZHANG JIE

Assignees

北京国网信通埃森哲信息技术有限公司

Dates

Publication Date: 20260508
Application Date: 20260123

Claims (10)

1. The dynamic privacy protection method for the electric power data based on the homomorphic encryption is characterized by comprising the following steps of: Acquiring power data to be processed, encrypting the power data by adopting an homomorphic encryption algorithm to obtain ciphertext power data, and establishing an initial noise reference value of the ciphertext power data; Constructing a ciphertext domain state evaluation model, analyzing dynamic association characteristics of ciphertext power data in a calculation process according to the ciphertext domain state evaluation model, predicting an executable depth boundary of a subsequent ciphertext operation, and generating a hierarchical privacy protection strategy containing a multistage noise threshold based on the executable depth boundary; Constructing a noise sensitivity curve for the ciphertext power data according to the hierarchical privacy protection strategy in a ciphertext domain, performing piecewise fitting calculation on the noise sensitivity curve to obtain noise threshold reference values of each level of operation, and performing dynamic comparison analysis on the noise threshold reference values and a predicted value of the ciphertext domain state evaluation model; And determining noise attenuation parameters of each level of operation according to the dynamic comparison analysis result, carrying out hierarchical optimization configuration on a ciphertext operation link based on the noise attenuation parameters, finishing decryption processing on the ciphertext operation result, and outputting a plaintext analysis result.
2. The method of claim 1, wherein encrypting the power data using a fully homomorphic encryption algorithm to obtain ciphertext power data and establishing an initial noise reference value for the ciphertext power data comprises: Performing data type identification and sensitivity classification on the electric power data, determining encryption strength requirements corresponding to different types of electric power data according to the sensitivity classification result, and selecting a key length and encryption parameter configuration scheme of a full homomorphic encryption algorithm based on the encryption strength requirements; and synchronously monitoring initial noise distribution characteristics introduced by the isomorphic encryption operation in the encryption process, and establishing an initial noise reference value of the ciphertext power data based on the mapping relation between the initial noise distribution characteristics and the sensitivity grading result.
3. The method of claim 1, wherein constructing a ciphertext domain state evaluation model, analyzing dynamic correlation characteristics of the ciphertext power data during computation according to the ciphertext domain state evaluation model, predicting an executable depth boundary of a subsequent ciphertext operation, and generating a hierarchical privacy protection policy comprising a multi-level noise threshold based on the executable depth boundary comprises: Decoupling ciphertext structure features and an initial noise reference value of ciphertext power data, constructing a ciphertext domain state evaluation model based on the ciphertext structure features and the initial noise reference value, and constructing a mapping function between a noise growth rate and a ciphertext operation type according to the ciphertext domain state evaluation model; parameterizing the ciphertext structure features according to the mapping function, mapping a ciphertext operation sequence to be executed to the ciphertext domain state evaluation model through depth feature reconstruction, and deducing a noise accumulation predicted value corresponding to each operation step in the ciphertext operation sequence by using the parameterized mapping function; based on the difference relation between the noise accumulation predicted value and a preset ciphertext decryptable noise upper limit, the executable depth boundary of the ciphertext operation sequence is remodelled by combining the ciphertext structure characteristics, the executable depth boundary is subjected to hierarchical progressive optimization according to noise sensitivity, and a corresponding noise threshold is set for each optimization level to form a hierarchical privacy protection strategy containing multistage noise thresholds.
4. The method of claim 3, wherein constructing a ciphertext domain state evaluation model based on the ciphertext structural features and the initial noise reference value, and wherein constructing a mapping function between a noise growth rate and a ciphertext operation type based on the ciphertext domain state evaluation model comprises: Performing dimension analysis and dependency relation extraction on the ciphertext structure features, obtaining hierarchical structure information and operation dependency graphs of ciphertext data, and performing hierarchical labeling on the initial noise reference value according to the hierarchical structure information to form a multi-level noise state of the ciphertext data; Classifying ciphertext operations based on operation type characteristics in the operation dependency graph according to the influence degree of operation operations, dividing the ciphertext operations into noise linear growth type operations and noise nonlinear growth type operations, and determining corresponding noise propagation parameters for each operation type; Inputting the multi-level noise state and the noise propagation parameter into a ciphertext domain state evaluation model, and tracking the noise evolution process of ciphertext in a continuous operation process through an operation path in the operation dependency graph according to the ciphertext domain state evaluation model; and carrying out fitting analysis on noise growth data in the noise evolution process based on the ciphertext domain state evaluation model, extracting noise growth rates corresponding to different ciphertext operations by combining the operation type characteristics, and establishing a mapping function between the noise growth rates and the ciphertext operations.
5. The method of claim 1, wherein constructing a noise sensitivity curve for the ciphertext power data in the ciphertext domain according to the hierarchical privacy preserving policy, and obtaining the noise threshold reference value for each stage of operation by performing a piecewise fitting calculation on the noise sensitivity curve comprises: Extracting a mapping relation between a multi-level noise threshold value and a noise tolerance interval from the hierarchical privacy protection strategy, performing a detective operation covering all the noise tolerance intervals on the ciphertext power data in a ciphertext domain, constructing a multi-dimensional operation path according to a combined link of the detective operation, monitoring the correlation characteristic of the noise evolution state and the operation depth of the ciphertext data in real time based on the multi-dimensional operation path, and constructing a noise sensitivity curve by utilizing the correlation characteristic; And carrying out derivative analysis on the noise sensitivity curve, identifying inflection point positions of noise growth acceleration mutation in the noise evolution state, dividing the noise sensitivity curve into a plurality of noise growth stage intervals by taking the inflection point positions as segmentation boundaries, extracting noise accumulation rate characteristics and curve curvature characteristics by combining the multidimensional operation path for each noise growth stage interval, and executing parameterization fitting calculation to deduce a noise threshold reference value of operation level of the corresponding stage interval.
6. The method of claim 5, wherein performing a exploratory operation on the ciphertext power data in a ciphertext domain that covers an entire noise margin interval, constructing a multi-dimensional operation path from a combined link of the exploratory operations comprises: Extracting boundary threshold values of all noise margin sections from a layered privacy protection strategy, designing a detection operation sequence covering all the noise margin sections according to the boundary threshold values, sequentially executing operation of increasing noise intensity on the ciphertext power data in a ciphertext domain based on the detection operation sequence, enabling ciphertext noise to pass through each noise margin section step by step from an initial noise state until reaching the maximum noise margin upper limit, and acquiring operation type identification, noise variation and noise margin section position of each operation in real time in the process, wherein a combination link of detection operation is formed based on a ternary combination relation of the operation type identification, the noise variation and the noise margin section position; And carrying out path analysis based on the combined link, extracting operation nodes with the same operation type identifier, dividing the operation nodes into stable propagation nodes and acceleration propagation nodes by analyzing the noise variation difference of the operation nodes in different noise margin intervals, and assembling the stable propagation nodes and the acceleration propagation nodes as path construction units according to the progressive sequence of the noise margin intervals to form a multidimensional operation path.
7. The method of claim 1, wherein determining noise attenuation parameters of each stage of operation according to the result of the dynamic comparison analysis, and performing hierarchical optimization configuration on ciphertext operation links based on the noise attenuation parameters, to complete decryption processing of the ciphertext operation result, and outputting a plaintext analysis result comprises: Performing piecewise fitting on noise variation characteristics in the ciphertext operation process according to the dynamic comparison analysis result, and determining noise attenuation parameters of each stage of operation; classifying ciphertext operation links based on the noise attenuation parameters, and dynamically configuring operation resources for different operation intervals according to the change trend of the noise attenuation parameter curve; And executing hierarchical optimization processing on the ciphertext operation link according to the configuration result of the operation resource, completing decryption operation on the ciphertext operation result, and outputting a plaintext analysis result.
8. A dynamic privacy protection system for electric power data based on isomorphic encryption, for implementing the method according to any of the previous claims 1-7, characterized in that it comprises: The encryption unit is used for acquiring the power data to be processed, carrying out encryption processing on the power data by adopting an homomorphic encryption algorithm to obtain ciphertext power data, and establishing an initial noise reference value of the ciphertext power data; The computing unit is used for constructing a ciphertext domain state evaluation model, analyzing dynamic association characteristics of the ciphertext power data in the computing process according to the ciphertext domain state evaluation model, predicting an executable depth boundary of a subsequent ciphertext operation, and generating a hierarchical privacy protection strategy containing a multistage noise threshold based on the executable depth boundary; the fitting unit is used for constructing a noise sensitivity curve for the ciphertext power data according to the hierarchical privacy protection strategy in a ciphertext domain, obtaining noise threshold reference values of each level of operation through carrying out sectional fitting calculation on the noise sensitivity curve, and carrying out dynamic comparison analysis on the noise threshold reference values and the predicted values of the ciphertext domain state evaluation model; and the configuration unit is used for determining noise attenuation parameters of each level of operation according to the dynamic comparison analysis result, carrying out hierarchical optimization configuration on the ciphertext operation link based on the noise attenuation parameters, completing decryption processing on the ciphertext operation result and outputting a plaintext analysis result.
9. An electronic device, comprising: A processor; A memory for storing processor-executable instructions; Wherein the processor is configured to invoke the instructions stored in the memory to perform the method of any of claims 1 to 7.
10. A computer readable storage medium having stored thereon computer program instructions, which when executed by a processor, implement the method of any of claims 1 to 7.

Description

Dynamic privacy protection method and system for electric power data based on isomorphic encryption Technical Field The invention relates to a data processing technology, in particular to a dynamic privacy protection method and system for electric power data based on isomorphic encryption. Background With the rapid development of smart grid technology, the amount of data generated and collected in a power system has been increasing in a burst. The power data contains sensitive information such as user electricity consumption behavior, power equipment running state, power grid load distribution and the like, and the safety and stability operation and user privacy protection of the power system are provided with serious challenges. The conventional power data processing method is mainly focused on the plaintext domain, which makes the user privacy information vulnerable to disclosure. The full homomorphic encryption technology is used as an advanced cryptography tool, allows calculation operation to be performed on encrypted data without decryption, and provides a new technical path for privacy protection of power data. The prior art generally adopts a static noise parameter configuration scheme, and cannot adapt to dynamic characteristic changes of power data in practical application. In the complex power calculation process, the data relevance and sensitivity can change along with the increase of the calculation depth, and the static noise configuration scheme is difficult to balance the dynamic relationship between the privacy protection intensity and the calculation efficiency. The application of the existing full homomorphic encryption in the power system ignores the influence of ciphertext operation depth on the quality of encrypted data. As the operation level increases, ciphertext noise gradually accumulates, and when a specific threshold is exceeded, decryption failure or result distortion is caused, and an effective prediction mechanism for an operation depth boundary is lacking, so that accuracy and reliability of power data analysis are affected. Disclosure of Invention The embodiment of the invention provides a dynamic privacy protection method and a dynamic privacy protection system for electric power data based on full homomorphic encryption, which can solve the problems in the prior art. In a first aspect of an embodiment of the present invention, there is provided a dynamic privacy protection method for electric power data based on isomorphic encryption, including: Acquiring power data to be processed, encrypting the power data by adopting an homomorphic encryption algorithm to obtain ciphertext power data, and establishing an initial noise reference value of the ciphertext power data; Constructing a ciphertext domain state evaluation model, analyzing dynamic association characteristics of ciphertext power data in a calculation process according to the ciphertext domain state evaluation model, predicting an executable depth boundary of a subsequent ciphertext operation, and generating a hierarchical privacy protection strategy containing a multistage noise threshold based on the executable depth boundary; Constructing a noise sensitivity curve for the ciphertext power data according to the hierarchical privacy protection strategy in a ciphertext domain, performing piecewise fitting calculation on the noise sensitivity curve to obtain noise threshold reference values of each level of operation, and performing dynamic comparison analysis on the noise threshold reference values and a predicted value of the ciphertext domain state evaluation model; And determining noise attenuation parameters of each level of operation according to the dynamic comparison analysis result, carrying out hierarchical optimization configuration on a ciphertext operation link based on the noise attenuation parameters, finishing decryption processing on the ciphertext operation result, and outputting a plaintext analysis result. Encrypting the power data by adopting an homomorphic encryption algorithm to obtain ciphertext power data, and establishing an initial noise reference value of the ciphertext power data comprises the following steps: Performing data type identification and sensitivity classification on the electric power data, determining encryption strength requirements corresponding to different types of electric power data according to the sensitivity classification result, and selecting a key length and encryption parameter configuration scheme of a full homomorphic encryption algorithm based on the encryption strength requirements; and synchronously monitoring initial noise distribution characteristics introduced by the isomorphic encryption operation in the encryption process, and establishing an initial noise reference value of the ciphertext power data based on the mapping relation between the initial noise distribution characteristics and the sensitivity grading result. Constructing a ciphertext domain