Nevra BAYHAN

Dr. Nevra BAYHAN

orcid

Research Focus & Metrics

Academic Career Length

19Years (2008-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
23 April 2024 in IEEE Access
Verified
Journal Article
Open access
YAYılmaz Seryar Arıkuşu
Yılmaz Seryar Arıkuşu
NBNevra BAYHAN
Nevra BAYHAN
The combined use of Extreme Gradient Boosting (XGBoost) algorithm, one of the machine learning (ML) methods, and a generalization of Hermite-Biehler theorem to obtain a novel PID controller that will ensure robustly stable and optimized operation of a micro thermoelectric cooler (Micro-TEC), which is the main part of a Polymerase chain reaction (PCR) device, is a unique approach of our study compared to previous studies. Therefore, we first established a mathematical model of the micro-TEC by making real-time measurements and then, a new data set was created to find the optimum parameter values of PID controller, and finally, XGBoost Hyperparameters with GridSearchCV was used for the first time to predict PID controller parameters. The XGBoost algorithm achieved 97% training success and 91% test success in estimating the parameters of the PID controller. Moreover, the novel controller developed using the XGBoost algorithm in this study has an impressive speed of 3 seconds. Additionally, our proposed method was compared with various metaheuristic optimization algorithms in terms of error percentage. The error percentages of XGBoost, the equilibrium optimization, the particle swarm optimization and the artificial bee colony optimization algorithms were found to be 0.4%, 1.1%, 3.7% and 11.1%, respectively. It is observed the settling times of micro-TEC with ML-PID controller for all five PCR cycles are 4.86, 44, 83.4, 123 and 162.5 seconds, respectively, and the overshoot values are below 5%. The proposed method gave the smallest settling time, error and overshoot percentages compared to these metaheuristic optimization algorithms.

8 December 2023 in Energies
Verified
Journal Article
Open access
PDF file
YAYılmaz Seryar Arıkuşu
Yılmaz Seryar Arıkuşu
NBNevra BAYHAN
Nevra BAYHAN
HTHasan TİRYAKİ
Hasan TİRYAKİ
In this study, a parallel hybrid electric vehicle produced within the scope of our project titled “Development of Fuel Efficiency Enhancing and Innovative Technologies for Internal Combustion Engine Vehicles” has been modeled. Firstly, a new rule-based control method is proposed to minimize fuel consumption and carbon emission values in driving cycles in the experimental model of the parallel hybrid electric vehicle produced within the scope of this project. The proposed control method ensures that the internal combustion engine (ICE) operates at the optimum point. In addition, the electric motor (EM) is activated more frequently at low speeds, and the electric motor can also work as a generator. Then, a new dataset was also created on a traffic-free racetrack with the proposed control method for fuel consumption estimation of a parallel hybrid electric vehicle using ECE-15 (Urban Driving Cycle), EUDC (Extra Urban Driving Cycle), and NEDC (New European Driving Cycle) driving cycles. The data set is dependent on 11 different input variables, which complicates the system. Afterward, the fuel estimation process is made with seven different machine learning methods (ML), and these methods are compared using the obtained data set. To avoid overfitting machine learning, two different test data sets were created. The Random Forest algorithm is the most suitable technique in terms of training and testing the fuel consumption model using correlation coefficient (R2), Mean Absolute Error (MAE), and Root Mean Square Error (RMSE) simulation appropriateness for both test datasets. Moreover, the random forest algorithm achieved an impressive accuracy of 97% and 90% for both test datasets, outperforming the other algorithms. Furthermore, the proposed method consumes 4.72 L of fuel per 100 km, while the gasoline-powered vehicle consumes 7 L of fuel per 100 km. The results show that the proposed method emits 4.69 kg less CO2 emissions. The effectiveness of the Random Forest Algorithm has been verified by both simulation results and real-world data.

22 September 2025 in International Journal of Refrigeration
Verified
Journal Article
Closed access
TKTufan Koç
Tufan Koç
NBNevra BAYHAN
Nevra BAYHAN
AYAminu Yusuf
Aminu Yusuf
Sedat BallikayaSedat Ballikaya
Sedat Ballikaya

19 December 2025 in Processes
Verified
Journal Article
Open access
PDF file
NBNevra BAYHAN
Nevra BAYHAN
YKYasin Koçak
Yasin Koçak
This study presents a systematic comparative analysis of nine stator current discretization methods within the Model Predictive Current Control (MPCC) framework for Permanent Magnet Synchronous Motors (PMSMs). These methods have generally been examined individually or in limited combinations in previous research, and this holistic and comprehensive comparison constitutes the core contribution of this work by addressing a significant gap in the existing literature. The investigated MPCC methods—Forward Euler (FE), Backward Euler (BE), Midpoint Euler (ME), Fourth-Order Runge–Kutta (RK4), Runge–Kutta Ralston (RKR), Taylor Series (TS), Verlet Integration (VI), Crank–Nicolson (CN), and Adams–Bashforth (AB)—are comprehensively evaluated for their dynamic performance, including speed tracking, torque response, settling time, rise time, overshoot, and Total Harmonic Distortion (THD). Additionally, these analysis results are benchmarked against conventional Proportional–Integral–Derivative (PID) and Field-Oriented Control (FOC) methods. In terms of key performance indicators, the MPCC–RKR method proved optimal for speed tracking under no-load conditions, achieving the lowest overshoot, specifically ranging from 0.097% to 1.450%. Conversely, MPCC–ME and MPCC–CN demonstrated superior transient performance under sudden-load conditions (1.7 Nm), yielding the smallest torque deviations, fastest settling times. Specifically, MPCC-ME recorded the lowest overshoot (1.512%) at the 7 s load step, while MPCC-CN performed best at 9 s (1.220%) and 11 s (1.577%). Among the predictive schemes, the MPCC–RKR method achieved the highest current quality with a minimum THD of 3.69% at nominal speed. Finally, it has been confirmed through the applied statistical analysis techniques that the performance differences among the discretization methods are significant. The comparative analysis examines both the dynamic performance of the methods and the fundamental trade-off between accuracy and computational burden in MPCC design. Simple single-step explicit methods (FE, ME, RKR, VI, AB) offer low computational cost and are well suited for high–sampling-frequency real-time applications, especially with sufficiently small sampling times, whereas more complex multi-step or implicit methods (BE, RK4, TS, CN) may increase the processor load despite their potential gains in accuracy and stability. This study provides practical, evidence-based guidelines for selecting an optimal discretization method by balancing accuracy and dynamic performance requirements for PMSM applications.

1 January 2008 in IFAC Proceedings Volumes
Verified
Journal Article
Closed access
PDF file
MSMehmet Turan Söylemez
Mehmet Turan Söylemez
NBNevra BAYHAN
Nevra BAYHAN

30 June 2021 in Bilecik Şeyh Edebali Üniversitesi Fen Bilimleri Dergisi
Verified
Journal Article
Closed access
PDF file
KSKunter Sercan Sezer
Kunter Sercan Sezer
NBNevra BAYHAN
Nevra BAYHAN
Bu çalışmada, klasik oransal-integral-türev (PID) kontrolörlerin gelişmiş hali olan kesir dereceli oransal-integral-türev (FOPID) kontrolörlerden faydalanılarak basitleştirilmiş gaz türbinli bir santral modeli için denetlemenin hassas şekilde yapılabilmesi sağlanmıştır. Klasik PID kontrolörler (denetleyiciler) üç parametre içerirken, kesir dereceli PID kontrolörler beş parametre içerir. Parametre sayısının fazla olması daha hassas denetlemenin yapılabilmesine olanak sağlar, ancak bu durum kontrolörün optimizasyonunu zorlaştırır. Kesir dereceli PID kontrolörlerin geleneksel matematiksel yöntemler ile optimizasyonu zor olduğu için; bu çalışmada, doğadan esinlenen (meta-sezgisel) optimizasyon algoritmaları arasında yer alan balina optimizasyon algoritması (BOA), salp sürüsü algoritması (SSA), yapay arı kolonisi (YAK) ve atom arama optimizasyon algoritması (AAO) kullanılmıştır. Bu dört farklı algoritmayla optimize edilen FOPID kontrolör parametreleri, basitleştirilmiş gaz türbini enerji santrali modeline uygulanmış ve sistem çıkış sinyallerinin geçici yanıt performansları karşılaştırılmıştır. Bu amaçla yerleşme süresi ve yüzde en büyük aşım, karşılaştırma kriteri olarak kullanılmıştır. Simulasyon sonuçları, yapay arı kolonisi (YAK) algoritmasıyla optimize edilmiş FOPID kontrolörün, bu santral modeli için yerleşme süresi ve yüzde en büyük aşım kriterleri açısından diğer algoritmalarla optimize edilmiş FOPID kontrolörlere göre daha iyi performans sergilediğini göstermektedir.