This paper presents a deep learning based multi-label attack detection approach for the distributed control in AC microgrids. The secondary control of AC microgrids is formulated as a constrained optimization problem with voltage and frequency as control variables which is then solved using a distributed primal-dual gradient algorithm. The normally distributed false data injection (FDI) attacks against the proposed distributed control are then designed for the distributed gener-ator's output voltage and active/reactive power measurements. In order to detect the presence of false measurements, a deep learning based attack detection strategy is further developed. The proposed attack detection is formulated as a multi-label classification problem to capture the inconsistency and co-occurrence dependencies in the power flow measurements due to the presence of FDI attacks. With this multi-label classification scheme, a single model is able to identify the presence of different attacks and load change simultaneously. Two different deep learning techniques are compared to design the attack detector, and the performance of the proposed distributed control and the attack detector is demonstrated through simulations on the modified IEEE 34-bus distribution test system.
more »
« less
This content will become publicly available on February 1, 2025
Detection of Multiplicative False Data Injection Cyberattacks on Process Control Systems via Randomized Control Mode Switching
A fundamental problem at the intersection of process control and operations is the design of detection schemes monitoring a process for cyberattacks using operational data. Multiplicative false data injection (FDI) attacks modify operational data with a multiplicative factor and could be designed to be detection evading without in-depth process knowledge. In a prior work, we presented a control mode switching strategy that enhances the detection of multiplicative FDI attacks in processes operating at steady state (when process states evolve within a small neighborhood of the steady state). Control mode switching on the attack-free process at steady-state may induce transients and generate false alarms in the detection scheme. To minimize false alarms, we subsequently developed a control mode switch-scheduling condition for processes with an invertible output matrix. In the current work, we utilize a reachable set-based detection scheme and use randomized control mode switches to augment attack detection capabilities. The detection scheme eliminates potential false alarms occurring from control mode switching, even for processes with a non-invertible output matrix, while the randomized switching helps bolster the confidentiality of the switching schedule, preventing the design of a detection-evading “smart” attack. We present two simulation examples to illustrate attack detection without false alarms, and the merits of randomized switching (compared with scheduled switching) for the detection of a smart attack.
more »
« less
- Award ID(s):
- 2137281
- NSF-PAR ID:
- 10509405
- Publisher / Repository:
- MDPI
- Date Published:
- Journal Name:
- Processes
- Volume:
- 12
- Issue:
- 2
- ISSN:
- 2227-9717
- Page Range / eLocation ID:
- 327
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
In a centralized Networked Control System (NCS), agents share local data with the central processing unit that generates control commands for agents. The control center in an NCS receives information from the agents through a communication network and produces control commands for agents. Despite all of the advantages of an NCS, such as reduced design cost and simplicity, the integration of networked connectivity can expose the NCS to adversarial attacks, such as false data injection (FDI). In this paper, a novel control approach will be developed to mitigate the FDI attack’s effect and guarantee the control objective in a networked system of permanent magnet linear motors. To achieve this, a non-singular terminal sliding mode control will be designed using an observer to ensure the tracking objective. The extended state observer will estimate the state of the system and estimate the FDI attack in real time. The control center will produce a control signal which is robust to the FDI attack and any disturbance. A Lyapunov-based stability analysis will be used to prove the stability of the observer-based controller. A three-agent permanent magnet linear motor network is selected for the simulation to show the effectiveness of the proposed scheme.more » « less
-
more » « less
Abstract The rise in smart water technologies has introduced new cybersecurity vulnerabilities for water infrastructures. However, the implications of cyber‐physical attacks on the systems like urban drainage systems remain underexplored. This research delves into this gap, introducing a method to quantify flood risks in the face of cyber‐physical threats. We apply this approach to a smart stormwater system—a real‐time controlled network of pond‐conduit configurations, fitted with water level detectors and gate regulators. Our focus is on a specific cyber‐physical threat: false data injection (FDI). In FDI attacks, adversaries introduce deceptive data that mimics legitimate system noises, evading detection. Our risk assessment incorporates factors like sensor noises and weather prediction uncertainties. Findings reveal that FDIs can amplify flood risks by feeding the control system false data, leading to erroneous outflow directives. Notably, FDI attacks can reshape flood risk dynamics across different storm intensities, accentuating flood risks during less severe but more frequent storms. This study offers valuable insights for strategizing investments in smart stormwater systems, keeping cyber‐physical threats in perspective. Furthermore, our risk quantification method can be extended to other water system networks, such as irrigation channels and multi‐reservoir systems, aiding in cyber‐defense planning.
-
This work focuses on the problem of enhancing cyberattack detection capabilities in process control systems subject to multiplicative cyberattacks. First, the relationship between closed-loop stability and attack detectability with respect to a class of residual-based detection schemes is rigorously analyzed. The results are used to identify a set of controller parameters (called "attack-sensitive" controller parameters) under which an attack can destabilize the closed-loop system. The selection of attack-sensitive controller parameters can enhance the ability to detect attacks, but can also degrade the performance of the attack-free closed-loop system. To balance this trade-off, a novel active attack detection methodology employing controller parameter switching between the nominal controller parameters (chosen on the basis of standard control design criteria) and the attack-sensitive controller parameters, is developed. The proposed methodology is applied to a chemical process example to demonstrate its ability to detect multiplicative sensor-controller communication link attacks.more » « less
-
Smart water metering (SWM) infrastructure collects real-time water usage data that is useful for automated billing, leak detection, and forecasting of peak periods. Cyber/physical attacks can lead to data falsification on water usage data. This paper proposes a learning approach that converts smart water meter data into a Pythagorean mean-based invariant that is highly stable under normal conditions but deviates under attacks. We show how adversaries can launch deductive or camouflage attacks in the SWM infrastructure to gain benefits and impact the water distribution utility. Then, we apply a two-tier approach of stateless and stateful detection, reducing false alarms without significantly sacrificing the attack detection rate. We validate our approach using real-world water usage data of 92 households in Alicante, Spain for varying attack scales and strengths and prove that our method limits the impact of undetected attacks and expected time between consecutive false alarms. Our results show that even for low-strength, low-scale deductive attacks, the model limits the impact of an undetected attack to only 0.2199375 pounds and for high-strength, low-scale camouflage attack, the impact of an undetected attack was limited to 1.434375 pounds.more » « less
