Real-Time Monitoring and Predictive Control Optimization for a 50 Kva Diesel Generator Using Generalized Predictive Controller
Year: 2025
Type: article
Abstract: Diesel generators remain essential in distributed power systems but are prone to voltage and frequency fluctuations, thermal instability, and efficiency loss under dynamic load conditions. This study develops a real-time monitoring and predictive optimization framework using the Generalized Predictive Controller (GPC) for a 50 kVA diesel generator to address these limitations. The research aims to maintain voltage and frequency stability, optimize thermal management, minimize overload risks, and enhance overall efficiency compared to conventional control techniques. The generator’s dynamic model was developed in MATLAB/Simulink, integrating subsystems for voltage regulation, thermal balance, overload analysis, and frequency control. Predictive models were constructed using state-space representations, with optimization based on the minimization of quadratic cost functions balancing state deviations and control effort. Simulation results show that GPC effectively maintained terminal voltage at 225 V and frequency at 50 Hz with minimal oscillation. The generator sustained an average efficiency between 85% and 95%, operating stably at 50 kW output for a 48 kW load without overloading. Thermal evaluation indicated controlled temperature within 75°C–80°C, despite variations in generated and dissipated heat. Voltage regulation remained within 4–9%, demonstrating robust control under load fluctuations. Comparative analysis revealed that GPC outperformed the Model Predictive Controller with Virtual Synchronous Machine (MPC-VSM), achieving an overall performance score of 72.3 compared to 63.9, representing a 13.1% improvement in efficiency and control stability. The study concludes that GPC provides superior adaptability and precision in diesel generator control, significantly improving energy utilization and mechanical reliability. It recommends extending the framework for hybrid renewable-diesel systems, where predictive control can enhance operational resilience, reduce fuel consumption, and support sustainable power generation in real-time applications.
Authors: Oyam-Jaja Tamunoipirinye, C.O. Ahiakwo., Hachimenum. N. Amadi
Language: English
Subfield: Control and Systems Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7106821652
FWCI:
Cites:
Cited by:
Related to: