SGID: A Semi-Synthetic Dataset for Injection Attacks in Smart Grid Systems

In recent years, smart grid systems have been widely deployed to improve the quality of electricity service provided to residential and industrial areas. Electricity smart metering and digital transfer can improve different types of immediate and long-term decisions being taken by electricity service providers. However, there are multiple security issues linked to smart grid systems. Malicious data injection into smart grid systems can negatively affect the smart grid system. This work aims to create a semi-synthetic dataset that includes real normal smart meter readings and multiple forms of malicious injected data(SGID); the injected data is synthetically created. The results show that different machine-learning models are able to detect intrusive data and distinguish it from normal data.