Maryam Kazerooni

Literature review on the applications of data mining in power systems

Power system is a highly interconnected network which delivers electric power to the electricity users. Sustaining the secure and reliable delivery of electric power requires continuous monitoring of the system. To process the large volumes of data obtained from the measuring devices, it is essential to investigate effective data enhancement techniques. In this paper, a comprehensive study on the applications of data mining in power systems is presented. Data visualization, clustering, outliers detection and classification are investigated as four major areas of data mining and related works in power systems which utilize each of these methods are presented in an organized and structured fashion.

Analysis of the grid harmonics and their impacts on distribution transformers

Widespread application of power electronics in industrial and residential loads has increased the harmonics of the grid significantly. The grid harmonics have destructive impacts on the gird components including distribution transformers. Studying the impacts of current harmonics on distribution transformers is crucial for grid design and maintenance. Therefore, this paper studies the effect of gird harmonics on eddy current loss, other stray losses, hottest spot temperature, and lifetime of distribution transformers. The current harmonics of six 100KVA transformers at a provincial power distribution center are measured in a one week test period and the impacts of grid harmonics on the investigated transformer are thoroughly discussed.

Estimation of geoelectric field for validating geomagnetic disturbance modeling

Geomagnetic induced currents (GICs) cause half cycle saturation of transformers and consequently increase the harmonic currents and reactive power in the grid. This paper focuses on the GIC modeling as an important geomagnetic disturbance (GMD) mitigation program. A linear model is presented which relates the transformer GICs to the geoelectric fields based on power network topology and conductances. Using the transformer GIC measurements, three effective techniques are proposed for estimating geoelectric fields, namely the Least-squares (LS) estimator, the Ridge Regression (RR) Estimator and the Lease Absolute Value (LAV) one. The effectiveness of the three estimators under various noise scenarios is validated for a small systems and a larger test case through simulation results.

Transmission System Geomagnetically Induced Current Model Validation

Geomagnetically induced currents (GICs) have the potential to severely disrupt power system operations by causing increased reactive power losses in the high-voltage transmission grid. This paper focuses on validating the model which relates the GICs to their deriving electric field. A novel validation framework is proposed with the advantage of considering the system uncertainties and network information availability. The effectiveness of this approaches is validated using both a 20 bus test system as well as actual GIC data provided by American Transmission Company.

Singular value decomposition in geomagnetically induced current validation

Geomagnetically induced currents (GICs) can cause transformer half cycle saturation and consequently affect the power grid with harmonic currents and increased reactive power loss. This paper focuses on using transformer neutral GIC measurements to help validate the GIC models. A linear GIC model is presented that relates the geoelectric field input to the transformer neutral GICs. To validate this model, a temporal GIC matrix model is introduced and the singular value decomposition (SVD) is used for estimating the geoelectric field. The effectiveness of the approach is shown using a 20 bus test case. Finally, the GIC model is extended to account for nonuniform electric field and its accuracy is evaluated through SVD.

Optimal load management of EV battery charging and optimization of harmonic impacts on distribution transformers

The growing concern about CO2 emissions and dependency on gasoline contributes to the increasing application of electric vehicles (EVs). EV battery chargers are non-linear loads and large-scale scale application of EVs increases the grid harmonics significantly. The grid harmonics have negative impacts on the components of the power system including distribution transformers. In this paper the impacts of EV charging on the load loss, hot-spot temperature and life time of distribution transformers are studied. Moreover, the EV charging load is formulated as an optimization problem and Karush-Kuhn-Tucker optimality conditions are employed for obtaining an optimal EV charging schedule. The EV load management program presented in this paper, maintains the safety operation of distribution transformers by minimizing their life time reduction.

Probabilistic modeling and reliability analysis for validating geomagnetically induced current data

Geomagnetically induced currents (GICs) cause half cycle saturation of transformers, which increases harmonic currents and reactive powers in the grid. This paper presents the least-squares (LS) estimation of the geoelectric field using the GIC measurements for model validation purposes. To account for more realistic noise scenarios, a general probabilistic GIC measurement model has been developed. Using the proposed model, the accuracy of the LS estimator and its reliability in assessing the geoelectric field are available analytically. The analytical results are verified through numerical simulations using a practical 20-bus test case.

Mitigation of Geomagnetically Induced Currents Using Corrective Line Switching

This paper considers line switching as a remedial action to mitigate the geomagnetically induced currents (GICs) in large-scale power syste-kazerooni-2753840ms. The algorithm uses linear sensitivity analysis to find the best switching strategy which minimizes the GIC-saturated reactive power loss. Furthermore, the coupling between the ac power flow solution and the GIC flows is modeled and heuristic strategies are designed to maintain the system security measures in terms of both GICs and ac power flow. Finally, the computational complexity of the algorithm is analyzed for large-system implementations. To reduce the complexity, the critical lines are identified through the sensitivity analysis and the calculations are performed only on the reduced system. The effectiveness of the proposed algorithm is demonstrated through numerical results using a small 20-bus test case as well as large power systems.

Incorporating the geomagnetic disturbance models into the existing power system test cases

Realistic public test cases can facilitate the studies on the GMDs impacts on the power system by providing a benchmark to validate the related analysis tools. Many standard test cases are available for different aspects of power system analysis. These cases are designed for ac analysis and do not contain the necessary inputs such as the substation grounding resistances and the geographic coordinates which are essential for GMD studies. In this paper, a framework is proposed to generate GMD-related parameters for the existing standard power system test cases. The substation geographic coordinates are the key parameters which are missing in the existing cases. The Kamada and Kawai algorithm and the Force-directed method are presented as two effective graph drawing algorithms to generate the geographic layout and determine the coordinates. The effectiveness of the proposed framework is evaluated through numerical results using the 20-bus system and the IEEE 24-bus system.

Distributed controller role and interaction discovery

Distributed controllers have a ubiquitous presence in the electric power grid and play a prominent role in its daily operation. The failure or malfunction of distributed controllers is a serious threat whose mechanisms and consequences are not currently well understood and planned against. For example, if certain controllers are maliciously compromised by an adversary, they can be manipulated to drive the power system to an unsafe state. We seek to develop proactive strategies to protect the power grid from distributed controller compromise or failure. This research formalizes the roles that distributed controllers play in the grid, quantifies how their loss or compromise impacts the system, and develops effective strategies for maintaining or regaining system control. Specifically, an analytic method based on controllability analysis is derived using clustering and factorization techniques on controller sensitivities.