Manufacturing Operations Synthetic Data Set for Anomaly/Cyberattack Detections

Cybersecurity has become increasingly crucial in manufacturing as factories and industrial systems become more interconnected and digitized.  High-profile incidents like Stuxnet have underscored the vulnerability of critical infrastructure to cyberattacks.  While machine learning (ML) has been successfully used for cyberattack detection in information technology (IT) environments, its application in operational technology (OT) environments, such as manufacturing, remains underexplored due to a lack of AI-ready datasets reflecting cyberattacks.  The lack of demonstration of ML application in OT environment limits our ability to quantify benefits of ML in OT environments and makes it challenging to justify investing in sensing capabilities to collect AI-ready OT datasets for experimentation and algorithm evaluation. To address this gap, we developed a generalizable OT data simulator to synthesize temporal manufacturing operations (MO) datasets under various cyberattack scenarios.

The "Manufacturing Operations Synthetic Data Set for Anomaly/Cyberattack Detections"  consists of simulations of a generic manufacutring process with statistical models consistent with nominal production (including production system degradation) and malicious attack data.   Please consult with README.md for the structure and dictionary of the data set.