Dataset of Noise Signals Generated by Smart Attackers for Disrupting State of Health and State of Charge Estimations in Battery Energy Storage Systems
Description
This dataset is derived from real-time MATLAB/Simscape simulations, focusing on the impact of subtle noise signals on battery energy storage systems. Using Proximal Policy Optimization (PPO), noise signals in millivolt and milliampere ranges are generated to stealthily disrupt the State of Charge (SoC) and State of Health (SoH) estimations within Unscented Kalman Filters (UKF). Designed to evade detection while causing estimation errors, this dataset is a valuable resource for studying and mitigating smart cyber-physical attacks. It can be reused in research to enhance the resilience of SoC and SoH estimation methods and develop robust defensive strategies.
Files
Steps to reproduce
1- Set Up the Simulation Environment: Access and download the Simscape Battery State-of-Health Estimation model from MathWorks: https://www.mathworks.com/help/simscape-battery/ug/battery-state-of-health-estimation.html Ensure the necessary MATLAB, Simulink, Simscape, and related toolboxes are installed. 2- Integrate the Noise Signals: Open the Simscape battery model in MATLAB. Identify the sensor blocks that measure current and voltage in the model. Inject the noise signals (out.currentFDIA for current and out.voltageFDIA for voltage) into the corresponding sensor outputs before they are fed into the SoC and SoH estimation blocks. 3-Run the Simulation: Execute the simulation with the noise signals integrated. Observe and log the impact on the SoC and SoH estimation values. 4- Analyze and Ensure Security: Evaluate how the noise affects the SoC and SoH estimates. Consider implementing countermeasures to enhance the security and robustness of the battery energy storage system.