Abstract
The measurement data in the power system may be attacked and tampered during the transmission. For this problem, a data encryption and approximate recovery strategy based on double random sorting and CGAN is proposed. The double random sorting encryption algorithm is proposed based on the form of measurement data in plaintext. The randomness of pseudo-random number, random sequence and random sequence insertion will ensure the uncertainty of data block number, sorting order and insertion location of random sequence, which will definitely improve the security of measurement data. Besides, an approximate recovery strategy used to approximately recovery abnormal samples is proposed on the base of CGAN, which can ensure the accuracy of decryption. Finally, the feasibility, and effectiveness of the proposed method are analyzed on wind power and photovoltaic power historical dataset.
Highlights
The rapid development of artificial intelligence algorithms and intelligent equipment promotes the development of the power system, which may lead to more potential vulnerabilities in the smart grid and bring great threats to the operation of the power system and the plaintext transmission of data
Data encryption technology has been widely used on preventing unknown attacks because of its wide applicability and simple structure. [3] proposed a micro encryption technology combining with the structure characteristics of IEC61850-9-2LE message, which realized the efficient encryption of data
We proposed an encryption method and an approximate recovery strategy based on double random sorting and conditional generative adversarial networks (CGAN), inspired by the algorithm of password based encryption (PBE)
Summary
The rapid development of artificial intelligence algorithms and intelligent equipment promotes the development of the power system, which may lead to more potential vulnerabilities in the smart grid and bring great threats to the operation of the power system and the plaintext transmission of data. Some real-time data in the power system is transmitted in plaintext form, like telemetering data. The wind power scenarios and PV power scenarios will be used as the input of day-ahead dispatching model [6] If these scenarios are tampered, the security of real-time operation of the power system will be affected. There are some shortcomings of the traditional encryption algorithms, such as complicated generation of key, so they are not suitable for the encryption of uncertain data. As for this problem, we proposed a data encryption and an approximate recovery strategy based on double random sorting and CGAN. The approximate data generated by trained CGAN are used to recovery the detected abnormal data in order to ensure the accuracy of the decrypted data
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have