Browsing by Author "Ceylan, O."

Now showing 1 - 6 of 6

Advanced Restoration Management Strategies in Smart Grids: the Role of Distributed Energy Resources and Load Priorities
(Elsevier Ltd, 2025) Ahmadi, B.; Ceylan, O.; Ozdemir, A.
Fast restoration following long outages is a challenge in the smart city management process. It is necessary to accurately characterize the real operating conditions of the system for optimal restoration. This study focuses on two key factors of a practical distribution system restoration. The first factor is cold load pickup (CLPU), which commonly occurs after an outage and is caused by thermostatically controlled loads. A time-dependent CLPU is modeled to accurately describe the restored load behaviors. The second factor is the effect of the distributed generators (DG), energy storage systems (ESSs), and load priority factors on the system's restoration process. To address this challenge, a robust optimization model is proposed that fully considers the effect of DG, and ESS units and uncertainty of CLPU. The proposed models are tested on the IEEE 33-node and 69-node test systems using the Advanced Grey Wolf Algorithm (AGWO). The simulation scenarios are designed to uncover optimal scheduling strategies for the restoration process corresponding to each Pareto solution of a previous study. The results are discussed for several distinct initial conditions. Moreover, a comparative evaluation is done, contrasting the outcomes achieved through the AGWO algorithm with those stemming from alternative heuristic methods. © 2025 The Authors
Citation - Scopus: 0
Data-Driven Methods for Optimal Setting of Legacy Control Devices in Distribution Grids
(IEEE Computer Society, 2024) Ceylan, Oğuzhan; Ceylan, O.; Paudyal, S.
This study presents machine learning-based dispatch strategies for legacy voltage regulation devices, i.e., onload tap changers (OLTCs), step-voltage regulators (SVRs), and switched-capacitors (SCs) in modern distribution networks. The proposed approach utilizes k-nearest neighbor (KNN), random forest (RF), and neural networks (NN) to map nodal net active and reactive injections to the optimal legacy controls and resulting voltage magnitudes. To implement these strategies, first, an efficient optimal power flow (OPF) is formulated as a mixed-integer linear program that obtains optimal decisions of tap positions for OLTCs, SVRs, and on/off status of SCs. Then, training and testing datasets are generated by solving the OPF model for daily horizons with 1-hr resolution for varying loading and photovoltaic (PV) generation profile. Case studies on the 33-node feeder demonstrate high-accuracy mapping between the input feature and the output vector, which is promising for integrated Volt/VAr control schemes. © 2024 IEEE.
Citation - Scopus: 2
Energy Loss Minimization with Parallel Implementation of Marine Predators Algorithm
(Institute of Electrical and Electronics Engineers Inc., 2021) Ceylan, Oğuzhan; Ahmadi, B.; Ceylan, O.; Ozdemir, A.
Distribution network (DN) service continuity is one of the significant issues in today's power systems. This paper aims to put a strategy for supplying loads with less discontinuity and affordable energy-consuming. The energy loss in distribution grids causes many problems for the producer and consumer; hence, it needs to be improved to increase supply efficiency accordingly. For this aim a model aiming to minimize power losses by allocating and sizing distributed generators (DGs) is solved using recently developed Marine Predators Algorithm (MPA). Since the proposed method is a time-intensive process due to the vast computations, parallel computation is implemented into MPA to increase computation speed. The proposed formulation and parallel computation are tested for 69-bus radial distribution system. The results are discussed in terms of computational accuracy and solution efficiency. Moreover, the convergence characteristics of MPA are compared with some other heuristic methods. © 2021 Chamber of Turkish Electrical Engineers.
Citation - Scopus: 0
Graph optimized locality preserving projection via heuristic optimization algorithms
(Institute of Electrical and Electronics Engineers Inc., 2019) Ceylan, Oğuzhan; Taşkin, G.
Dimensionality reduction has been an active research topic in hyperspectral image analysis due to complexity and nonlinearity of the hundreds of the spectral bands. Locality preserving projection (LPP) is a linear extension of the manifold learning and has been very effective in dimensionality reduction compared to linear methods. However, its performance heavily depends on construction of the graph affinity matrix, which has two parameters need to be optimized: k-nearest neighbor parameter and heat kernel parameter. These two parameters might be optimally chosen simply based on a grid search when using only one representative kernel parameter for all the features, but this solution is not feasible when considering a generalized heat kernel in construction the affinity matrix. In this paper, we propose to use heuristic methods, including harmony search (HS) and particle swarm optimization (PSO), in exploring the effects of the heat kernel parameters on embedding quality in terms of classification accuracy. The preliminary results obtained with the experiments on the hyperspectral images showed that HS performs better than PSO, and the heat kernel with multiple parameters achieves better performance than the isotropic kernel with single parameter. © 2019 IEEE.
Citation - Scopus: 0
Optimization of Graph Affinity Matrix With Heuristic Methods in Dimensionality Reduction of Hypespectral Images
(Institute of Electrical and Electronics Engineers Inc., 2019) Ceylan, O.; Ceylan, Oğuzhan; Taskin, G.
Hyperspectral images include hundreds of spectral bands, adjacent ones of which are often highly correlated and noisy, leading to a decrease in classification performance as well as a high increase in computational time. Dimensionality reduction techniques, especially the nonlinear ones, are very effective tools to solve these issues. Locality preserving projection (LPP) is one of those graph based methods providing a better representation of the high dimensional data in the low-dimensional space compared to linear methods. However, its performance heavily depends on the parameters of the affinity matrix, that are k-nearest neighbor and heat kernel parameters. Using simple methods like grid-search, optimization of these parameters becomes very computationally demanding process especially when considering a generalized heat kernel, including an exclusive parameter per feature in the high dimensional space. The aim of this paper is to show the effectiveness of the heuristic methods, including harmony search (HS) and particle swarm optimization (PSO), in graph affinity optimization constructed with a generalized heat kernel. The preliminary results obtained with the experiments on the hyperspectral images showed that HS performs better than PSO, and the heat kernel with multiple parameters achieves better performance than the heat kernel with a single parameter. © 2019 IEEE.
Citation - Scopus: 0
Synthetic Data for Non-Intrusive Load Monitoring: a Markov Chain Based Approach
(Institute of Electrical and Electronics Engineers Inc., 2024) Sayilar, B.C.; Mihci, G.; Ceylan, O.
This paper deals with the generation of synthetic data, which plays an important role in the Non-Intrusive Load Monitoring (NILM) problem. We introduce the NILM problem and then explain its crucial role in improving energy efficiency and supporting smart grid functions. The paper explains the stages of the NILM problem, including data acquisition, feature extraction, event detection, and appliance classification. We also explain two methods for generating synthetic data: AMBAL (Appliance Model Based Algorithm for Load monitoring) and SmartSim. Then, we propose a synthetic data generation method based on Markov chains, which is designed to generate labeled data useful for training supervised machine learning models. The proposed method utilizes the probabilistic transitions between different operational states of appliances, and captures the stochastic nature of real-world appliance usage. Thus, the generated synthetic data not only reflects realistic usage patterns, but also contains labels indicating the state of each appliance at a given time. The simulations are then run by generating synthetic data for typical office equipment such as laptops and televisions. The generated data sets provide detailed and accurate usage profiles, which are important for the effective training and validation of NILM algorithms. Since the generated data also includes the labeled data, this method will improve the ability of NILM systems to accurately identify and monitor individual appliances in a complex load environment. © 2024 IEEE.