Scientists of the St. Petersburg Federal Research Center of the Russian Academy of Sciences (SPC RAS) have developed a program that ensures a cybersecurity of automated water treatment monitoring systems at industrial enterprises and municipal services. For the program’s “learning” the scientists have developed the first in Russia training stand to acquire and monitor data from wastewater treatment plants operating under regular conditions and under cyberattacks.
Water is one of the basic needs of humanity. When it comes to industrial enterprises or densely inhabited big cities, the most important role is played by systems monitoring the quality of the water used by humans for consumption and household needs.
In recent years, the infrastructures for water treatment and water purification have been actively automated, for instance, remote access to control and monitoring systems has appeared. Same time, there exists a possibility of intruders’ cyberattacks against such infrastructures. Therefore, detection of various violations in the operation of automated water purification systems is an important task intended for taking timely measures and ensuring safety for public health.
“We have developed a program that, using the methods of explicable artificial intelligence, allows for monitoring the malfunction of automated water treatment plant systems. First and foremost, such abnormal activity is associated with the implementation of various cyberattack scenarios,” tells Elena Fedorchenko, senior researcher at the Laboratory of computer security problems of SPC RAS.
Program is based on an artificial neural network capable of distinguishing the normal operation of wastewater treatment plants from its functioning under a cyberattack by the parameters of signals from sensors of automated systems.
To teach the neural network to correctly identify violations in the wastewater disposal system, scientists of SPC RAS developed the first domestic semi-natural stand modeling the processes of automation of water purification; it can be adapted to water treatment and sanitation systems of various organizations.
“Due to the use of the mentioned stand, we can teach a neural network to detect a variety of cyberattack types, as well as provide protection measures already at the stage of implementing the automation systems. Our development will find application at institutions and companies engaged in water treatment by the nature of their activities, for instance, industrial enterprises or urban water utilities,” explains Elena Fedorchenko.
The project was supported by a grant from RSF and SPbSF (No. 23-11-20024).
The results of the study are partially published in a scientific journal