arXiv subjects

Ali Mehrizi-Sani

Publications and source records attributed to Ali Mehrizi-Sani.

11 recordsLinked to original sources

arXiv2024

Hybrid Machine Learning Approach for Cyberattack Mitigation of Parallel Converters in a DC Microgrid

Cyberattack susceptibilities are introduced as the communication requirement increases with the incorporation of more renewable energy sources into DC microgrids. Parallel DC-DC converters are utilized to provide high current and supply the load. Nevertheless, these systems are susceptible to cyberattacks that have the potential to disrupt operations and jeopardize stability. Voltage instability may result from the manipulation of communication commands and low-layer control signals. Therefore, in this paper, a cyberattack that specifically targets parallel DC-DC converters is examined in a DC microgrid. A hybrid machine learning-based detection and mitigation strategy is suggested as a means to counteract this threat. The false data injection (FDI) attack targeting the converters is investigated within a DC microgrid. The efficacy of the suggested approach is verified via simulations executed for various scenarios within the MATLAB/Simulink environment. The technique successfully identifies and blocks FDI attacks, preventing cyberattacks and ensuring the safe operation of the DC microgrid.

eess.SY↗

arXiv2024

Accurate Current Sharing in a DC Microgrid Using Modified Droop Control Algorithm

Due to the increasing popularity of DC loads and the potential for higher efficiency, DC microgrids are gaining significant attention. DC microgrids utilize multiple parallel converters to deliver sufficient power to the load. However, a key challenge arises when connecting these converters to a common DC bus: maintaining voltage regulation and accurate current sharing. Unequal cable resistances can cause uneven power sharing and lead to power losses. Conventional droop control methods, which employ a virtual resistor to address this issue, have limitations in achieving good performance across the entire converter operating range. This paper proposes a modified droop control algorithm to address this issue. This method modifies the virtual resistor in a way that ensures power sharing aligns with each converter-rated capacity. The algorithm is simple to implement and uses local measurements to update the droop gain. This paper presents simulation studies and experimental tests to analyze the performance of the proposed method, considering scenarios with equal and unequal converter ratings. The results successfully validate the accuracy and effectiveness of this innovative approach.

eess.SY↗

arXiv2024

Machine Learning-Based Protection and Fault Identification of 100% Inverter-Based Microgrids

100% inverter-based renewable units are becoming more prevalent, introducing new challenges in the protection of microgrids that incorporate these resources. This is particularly due to low fault currents and bidirectional flows. Previous work has studied the protection of microgrids with high penetration of inverter-interfaced distributed generators; however, very few have studied the protection of a 100% inverter-based microgrid. This work proposes machine learning (ML)-based protection solutions using local electrical measurements that consider implementation challenges and effectively combine short-circuit fault detection and type identification. A decision tree method is used to analyze a wide range of fault scenarios. PSCAD/EMTDC simulation environment is used to create a dataset for training and testing the proposed method. The effectiveness of the proposed methods is examined under seven distinct fault types, each featuring varying fault resistance, in a 100% inverter-based microgrid consisting of four inverters.

eess.SY↗

arXiv2024

Real-Time Simulation of a Resilient Control Center for Inverter-Based Microgrids

The number of installed remote terminal units (RTU) is on the rise, increasing the observability and control of the power system. RTUs enable sending data to and receiving data from a control center in the power system. A distribution grid control center runs distribution management system (DMS) algorithms, where the DMS takes control actions during transients and outages, such as tripping a circuit breaker and disconnecting a controllable load to increase the resiliency of the grid. Relying on communication-based devices makes the control center vulnerable to cyberattacks, and attackers can send falsified data to the control center to cause disturbances or power outages. Previous work has conducted research on developing ways to detect a cyberattack and ways to mitigate the adverse effects of the attack. This work studies false data injection (FDI) attacks on the DMS algorithm of a fully inverter-based microgrid in real time. The fully inverter-based microgrid is simulated using an RTDS, an amplifier, an electronic load, a server, a network switch, and a router. The DMS is integrated into the server codes and exchanges data with RTDS through TCP/IP protocols. Moreover, a recurrent neural network (RNN) algorithm is used to detect and mitigate the cyberattack. The effectiveness of the detection and mitigation algorithm is tested under various scenarios using the real-time testbed.

eess.SY↗

arXiv2024

Memory-Based Set Point Modulation for Improved Transient Response of Distributed Energy Resources

As the composition of the power grid evolves to integrate more renewable generation, its reliance on distributed energy resources (DER) is increasing. Existing DERs are often controlled with proportional integral (PI) controllers that, if not properly tuned or if system parameters change, exhibit sluggish performance or large overshoot. The use of set point automatic adjustment with correction-enabled (SPAACE) with a linear predictor improves the transient response of these DERs without the need to access the PI controller parameters. The limitation of the existing SPAACE method is the high sampling rate needed for improved performance, which is not always practical. This paper proposes the addition of a memory term to the SPAACE with a linear predictor. This memory term is the integral of the errors of previous samples, which adds another layer to the prediction to improve the response at lower sampling rates and further reduces the overshoot and settling time compared to the existing SPAACE method. Time-domain simulation studies are performed in PSCAD/EMTDC to show the effectiveness of the proposed controller.

eess.SY↗

arXiv2024

Stability Enhancement of LCL-Type Grid-Following Inverters Using Capacitor Voltage Active Damping

An LCL filter offers superior attenuation for high-frequency harmonics for three-phase grid-following inverters compared to LC and L filters. However, it also introduces an inherent resonance peak, which can lead to power quality issues or even instability of the inverter control system. Active damping (AD) is widely employed to effectively mitigate this resonance. Capacitor voltage feedback (CVF) and capacitor current feedback (CCF) are effective AD methods for LCL resonance damping. CVF is preferred due to its lower sensor requirement compared to CCF. However, a derivative term appears in the active damping loop, which introduces high-frequency noise into the system. This paper proposes a noise-immune approach by replacing the derivative term with a discrete function suitable for digital implementation. The LCL resonance can be damped effectively, resulting in enhanced stability of the inverter control system. Simulation results verify the proposed effectiveness of the method with grid inductance variation and weak grid conditions

eess.SY↗

arXiv2024

Detection and Mitigation of Cyberattacks on Volt-Var Control

Cyberattacks are becoming more frequent, and attackers can use different mechanisms, such as denial of service (DoS) and false data injection (FDI). Furthermore, multiple attack types can be launched simultaneously, known as hybrid attacks, to cause more damage. Volt-Var control algorithms are widely used in the distribution system to maintain the voltage within a nominal range. This work uses an artificial neural network (ANN)-based method to detect and mitigate hybrid cyberattacks on the Volt-VAr control algorithm.

eess.SY↗

arXiv2023

Scalable Optimal Design of Incremental Volt/VAR Control using Deep Neural Networks

Volt/VAR control rules facilitate the autonomous operation of distributed energy resources (DER) to regulate voltage in power distribution grids. According to non-incremental control rules, such as the one mandated by the IEEE Standard 1547, the reactive power setpoint of each DER is computed as a piecewise-linear curve of the local voltage. However, the slopes of such curves are upper-bounded to ensure stability. On the other hand, incremental rules add a memory term into the setpoint update, rendering them universally stable. They can thus attain enhanced steady-state voltage profiles. Optimal rule design (ORD) for incremental rules can be formulated as a bilevel program. We put forth a scalable solution by reformulating ORD as training a deep neural network (DNN). This DNN emulates the Volt/VAR dynamics for incremental rules derived as iterations of proximal gradient descent (PGD). Analytical findings and numerical tests corroborate that the proposed ORD solution can be neatly adapted to single/multi-phase feeders.

math.OC↗

arXiv2022

Reinforcement Learning$\unicode{x2013}$Based Transient Response Shaping for Microgrids

This work explores the usage of a supplementary controller for improving the transient performance of inverter$\unicode{x2013}$based resources (IBR) in microgrids. The supplementary controller is trained using a reinforcement learning (RL)$\unicode{x2013}$based algorithm to minimize transients in a power converter connected to a microgrid. The controller works autonomously to issue adaptive, intermediate set points based on the current state and trajectory of the observed or tracked variable. The ability of the designed controller to mitigate transients is verified on a medium voltage test system using PSCAD/EMTDC.

eess.SY↗

arXiv2022

Integrated Fault Diagnosis and Control Design for DER Inverters using Machine Learning Methods

This paper employs a supervised machine learning (ML) algorithm to propose an integrated fault detection and diagnosis (FDD) and fault-tolerant control (FTC) strategy to detect, diagnose, and classify the grid faults and correct the input voltage before affecting the grid-connected distributed energy resources (DER) inverters. This controller can mitigate the impact of grid faults on inverters by predicting and modifying the time series of their input voltage. Simulation results show the effectiveness of the proposed controller and evaluate its operating performance.

eess.SY↗

arXiv2021

Power Systems Performance under 5G Radio Access Network in a Co-Simulation Environment

Communication can improve control of important system parameters by allowing different grid components to communicate their states with each other. This information exchange requires a reliable and fast communication infrastructure. 5G communication can be a viable means to achieve this objective. This paper investigates the performance of several smart grid applications under a 5G radio access network. Different scenarios including set point changes and transients are evaluated, and the results indicate that the system maintains stability when a 5Gnetwork is used to communicate system states.

cs.NI↗