Τεχνικές ανίχνευσης και απόκρισης κυβερνοαπειλών για ηλεκτρικά οχήματα εξωτερικής τροφοδοσίας
Abstract
One of the biggest problems of modern society is considered the air pollution, which has led to the discovery of alternative energy sources in combination with other factors, such as the reduction of natural resources due to their unbridled and uncontrolled exploitation. The interim solution of using catalysts to reduce pollution proved insuffi-cient, because it simply limited the problem without leading to its final solution. The electric vehicle ensures zero emissions and frees users from dependence on liquid fuels, the sharp rise in prices and all kinds of shortages due to crises.
The various vehicles, on a case-by-case basis, are equipped with the correspond-ing fuels to power their engines from each of the refueling stations. Therefore, to charge electric vehicles, charging point stations are needed to supply electricity. Their supply, as mentioned, requires their connection to some kind of electricity network infrastructure. The large area of the electricity grid offers many options for potential charging facilities.
This thesis examines threats and cyber-attacks in the charging process and the tar-geting Battery Management System (BMS) and other parts of the car‘s electronics‘ sys-tem. More specifically, this thesis will examine whether the charging station hardware can be hacked in order to send these erroneous signals (either locally or remotely) and how the charging stations can be made tamper-proof and how cyber-attacks can be de-tected.
Charging Point Stations have many functions, such as, providing and controlling the energy to the Electric Vehicle (EV) using the Electric Vehicle Supply Equipment (EVSE) component, collecting the measurements from the meter for each charge of an Electric Vehicle, identifying and authorizing EV users via user authentication component, enabling remote capabilities (e.g., adjustment of the maximum current allowed by the Charge Point) to the Charge Point via the local Controller component over the Wide Area Network (WAN).
The protection of the European electric grid should become a priority for all the organization/entities that are getting engaged in the EV ecosystem. The output of this the-sis is aiming at increasing the cyber security of a standard EV charging enterprise‘s plat-form through the integration of Machine Learning (ML) techniques for identifying anom-alies in the charging patterns, and therefore minimize the exposure both enterprises‘ da-tabase and the stability of the electric grid. The thesis covers both the Information and Communications Technology (ICT) and the electric engineering domain on an effort to-wards increasing the cyber security on what is called Energy Internet.
In the implementation part of this thesis, we will use dataset in CSV format ob-tained from a standard EV charging enterprise‘s database to apply anomaly based algo-rithm, in order to discover if any abnormal functions of charging happens. For the smart charging abuse scenario, different evaluation methods will be applied in order to ensure high quality to the findings of the ML techniques. The applied evaluation methods will contain qualitative (visual inspection, manual investigation) metrics offering a validation framework wide enough to cover different aspects of cyber security in the area of EV smart charging.