J. Hill, J. Ye, W. Song, L. Lu
Aging Aircraft Solutions & University of Georgia Center for Cyberphysical Systems,
United States
Keywords: electrical waveforms, cybersecurity, manufacturing
Summary:
Electrical Waveform Fault & Cyberattack Detection System (EWFADS) monitors the waveforms of electricity powering manufacturing machines. Our research found that electrical waveforms offer unique signatures in different condition scenarios that can be used to identify and diagnose faults and attacks. The EWFADS solution will be capable of monitoring the condition of manufacturing machines with minimal integration or interference with machine hardware or software; the only interface will be on the electrical power input. This dramatically lowers the implementation cost, integration cost, and cybersecurity risks of this machine condition monitoring solution in compared with other condition monitoring solutions. Research found that electrical waveforms offer unique signatures in different condition scenarios that can be used to identify and diagnose faults and attacks using a random forest machine learning classifier. A classic cyberattack on a manufacturing machine could be a false data injection (FDI) attack. An FDI is a malicious attack that compromises the integrity of sensor readings inside the manufacturing machine to disrupt or “trick” the controller state estimation and outputs. Stuxnet is probably the most famous example of this type of attack on an industrial machine. Our research to date has shown to detect this type of attack.