Diseño de un controlador adaptativo para el control de nivel de tres tanques interconectados bajo ataques de inyección de retardos en el sistema de control industrial

Abstract

The objective of this work is to design an adaptive controller to maintain the level of a three-tank interconnected system under delay injection attacks. The proposal consists of estimating the delay introduced by the attacker and using this information to adapt the Proportional Integral Derivative (PID) controller to mitigate the effects caused by the attacker and, thus, allow the system to operate normally. The development of the work consisted of the mathematical modeling of the three-tank plant with its respective PID control system in the SIMULINK environment. To emulate the attacking injections, time-delay injections into the feedback and feedforward channels were simulated. For the delay estimation, the Grey-Box system identification technique using the Recursive Least Squares Method (RLSM) was employed. Finally, an adaptive PID controller was designed and implemented, whose gains are adjusted based on the delay estimation. The obtained results showed that time-delay injection attacks severely affect the performance of a conventional PID controller, leading the system to instability. In contrast, the adaptive PID controller that was used successfully mitigated the attack, preserving system stability for up to 200 seconds, even under attack conditions. The comparative study corroborated the greater robustness of the adaptive scheme. It is concluded that the insertion of an estimation mechanism together with an adaptive control strategy is a suitable solution for defending Industrial Control Systems (ICS) against cyber threats such as time-delay injection. This proposal enhances the resilience of ICS, thereby increasing the operability of critical infrastructures by offering availability, security, and reliability. Keywords: Adaptive Controller, Cybersecurity, Industrial Control Systems, Delay Injection, Parameter Estimation.