Othman Maganga - Academia.edu (original) (raw)

Papers by Othman Maganga

Research paper thumbnail of Model complexity reduction of a DC-DC buck-boost converter

The paper presents development of a reduced complexity model of a DC-DC buck- boost converter usi... more The paper presents development of a reduced complexity model of a DC-DC buck- boost converter using a methodology driven by system identification. A detailed high fidelity two-input single-output white-box model of the DC-DC buck-boost converter implemented in Simulink/Simscape is approximated by a set of relatively simple single-input single-output switched Hammerstein systems that are obtained at each of the 10 levels of the input voltage. To ensure a smooth transition between separate models the switching is realised by using a Gaussian blending approach. The results obtained show high modelling accuracy of the identified reduced complexity model.

Research paper thumbnail of Modelling and simulation of a thermoelectric generator for waste heat energy recovery in Low Carbon Vehicles

2012 2nd International Symposium On Environment Friendly Energies And Applications, 2012

ABSTRACT This paper details the development of a thermoelectric generator (TEG) model undertaken ... more ABSTRACT This paper details the development of a thermoelectric generator (TEG) model undertaken within the Low Carbon Vehicle Technology Project (LCVTP*). The model has been developed as a tool for investigating the application of using a TEG for waste energy recovery from an automotive engine exhaust and converting it to electrical energy to offset the electrical demand on the 12V battery. The model features three main subsystems that make up the TEG system; these are the heat exchanger, the thermoelectric material and the power conditioning unit. Particular attention has been given to the power conditioning unit where two different DC-DC converter topologies namely; buck-boost and single ended primary inductor capacitor (SEPIC) have been compared for best performance. In addition, the Perturb and Observe maximum power point tracking algorithm has also been implemented and compared with a standard fixed duty cycle pulse width modulator (PWM) control. The process of developing the subsystems are briefly explained and the advantages of using the maximum power point system is demonstrated. The simulation results demonstrate that a power conditioning unit with a buck-boost converter and Perturb and Observe control is suitable for TEG systems. MATLAB/Simulink has been used for modelling and simulation of the system as well as implementation of the control strategy

Research paper thumbnail of Investigation of Model Order Reduction Techniques: A Supercapacitor Case Study

Advances in Intelligent Systems and Computing, 2014

ABSTRACT This paper presents several different model order reduction techniques to refine an equi... more ABSTRACT This paper presents several different model order reduction techniques to refine an equivalent circuit high order model of a supercapacitor. The presented model order reduction techniques are: truncation based, projection based and system identification based (data based). Upon application of these techniques to the high order model, it has been found that a reduced model with sufficient accuracy can be obtained to act as a surrogate of the real system. This is evident by the ability to reduce a 60th order supercapacitor model to 4th order whilst preserving accuracy.

Research paper thumbnail of Hardware Implementation of Maximum Power Point Tracking for Thermoelectric Generators

Journal of Electronic Materials, 2014

ABSTRACT This work describes the practical implementation of two maximum power point tracking (MP... more ABSTRACT This work describes the practical implementation of two maximum power point tracking (MPPT) algorithms, namely those of perturb and observe, and extremum seeking control. The proprietary dSPACE system is used to perform hardware in the loop (HIL) simulation whereby the two control algorithms are implemented using the MATLAB/Simulink (Mathworks, Natick, MA) software environment in order to control a synchronous buck–boost converter connected to two commercial thermoelectric modules. The process of performing HIL simulation using dSPACE is discussed, and a comparison between experimental and simulated results is highlighted. The experimental results demonstrate the validity of the two MPPT algorithms, and in conclusion the benefits and limitations of real-time implementation of MPPT controllers using dSPACE are discussed.

Research paper thumbnail of Investigation of Maximum Power Point Tracking for Thermoelectric Generators

Journal of Electronic Materials, 2013

ABSTRACT In this paper, a thermoelectric generator (TEG) model is developed as a tool for investi... more ABSTRACT In this paper, a thermoelectric generator (TEG) model is developed as a tool for investigating optimized maximum power point tracking (MPPT) algorithms for TEG systems within automotive exhaust heat energy recovery applications. The model comprises three main subsystems that make up the TEG system: the heat exchanger, thermoelectric material, and power conditioning unit (PCU). In this study, two MPPT algorithms known as the perturb and observe (P&O) algorithm and extremum seeking control (ESC) are investigated. A synchronous buck–boost converter is implemented as the preferred DC–DC converter topology, and together with the MPPT algorithm completes the PCU architecture. The process of developing the subsystems is discussed, and the advantage of using the MPPT controller is demonstrated. The simulation results demonstrate that the ESC algorithm implemented in combination with a synchronous buck–boost converter achieves favorable power outputs for TEG systems. The appropriateness is by virtue of greater responsiveness to changes in the system’s thermal conditions and hence the electrical potential difference generated in comparison with the P&O algorithm. The MATLAB/Simulink environment is used for simulation of the TEG system and comparison of the investigated control strategies.

Research paper thumbnail of A comparative Study of Model-Based and Data-Based Model Order Reduction Techniques for Nonlinear Systems

Advances in Intelligent Systems and Computing, 2014

New technologies and multi-physical description of subsystems have forced designers to consider n... more New technologies and multi-physical description of subsystems have forced designers to consider nonlinear effects for more accurate modelling leading to increased complexity of mathematical models . Such complex models are non-trival to analyse and to develop control algorithms. Consequently, increasing complexity of circuit designs causes the need for model order reduction (MOR) techniques that are capable of reducing nonlinear models and decreasing computational cost of simulating nonlinear systems. MOR techniques for linear time invariant (LTI) systems are well established [2]. On the other hand MOR for nonlinear systems is an open problem [1]. There are several ways of obtaining reduced order model (ROM) for nonlinear systems via model-based approach such as linear approximation(LA) [3], bilinearisation, proper orthogonal decomposition (POD), quadratic approximation (QA) and trajectory piecewise linear (TPWL) approximation, etc.

Research paper thumbnail of Model complexity reduction of a DC-DC buck-boost converter

The paper presents development of a reduced complexity model of a DC-DC buck- boost converter usi... more The paper presents development of a reduced complexity model of a DC-DC buck- boost converter using a methodology driven by system identification. A detailed high fidelity two-input single-output white-box model of the DC-DC buck-boost converter implemented in Simulink/Simscape is approximated by a set of relatively simple single-input single-output switched Hammerstein systems that are obtained at each of the 10 levels of the input voltage. To ensure a smooth transition between separate models the switching is realised by using a Gaussian blending approach. The results obtained show high modelling accuracy of the identified reduced complexity model.

Research paper thumbnail of Modelling and simulation of a thermoelectric generator for waste heat energy recovery in Low Carbon Vehicles

2012 2nd International Symposium On Environment Friendly Energies And Applications, 2012

ABSTRACT This paper details the development of a thermoelectric generator (TEG) model undertaken ... more ABSTRACT This paper details the development of a thermoelectric generator (TEG) model undertaken within the Low Carbon Vehicle Technology Project (LCVTP*). The model has been developed as a tool for investigating the application of using a TEG for waste energy recovery from an automotive engine exhaust and converting it to electrical energy to offset the electrical demand on the 12V battery. The model features three main subsystems that make up the TEG system; these are the heat exchanger, the thermoelectric material and the power conditioning unit. Particular attention has been given to the power conditioning unit where two different DC-DC converter topologies namely; buck-boost and single ended primary inductor capacitor (SEPIC) have been compared for best performance. In addition, the Perturb and Observe maximum power point tracking algorithm has also been implemented and compared with a standard fixed duty cycle pulse width modulator (PWM) control. The process of developing the subsystems are briefly explained and the advantages of using the maximum power point system is demonstrated. The simulation results demonstrate that a power conditioning unit with a buck-boost converter and Perturb and Observe control is suitable for TEG systems. MATLAB/Simulink has been used for modelling and simulation of the system as well as implementation of the control strategy

Research paper thumbnail of Investigation of Model Order Reduction Techniques: A Supercapacitor Case Study

Advances in Intelligent Systems and Computing, 2014

ABSTRACT This paper presents several different model order reduction techniques to refine an equi... more ABSTRACT This paper presents several different model order reduction techniques to refine an equivalent circuit high order model of a supercapacitor. The presented model order reduction techniques are: truncation based, projection based and system identification based (data based). Upon application of these techniques to the high order model, it has been found that a reduced model with sufficient accuracy can be obtained to act as a surrogate of the real system. This is evident by the ability to reduce a 60th order supercapacitor model to 4th order whilst preserving accuracy.

Research paper thumbnail of Hardware Implementation of Maximum Power Point Tracking for Thermoelectric Generators

Journal of Electronic Materials, 2014

ABSTRACT This work describes the practical implementation of two maximum power point tracking (MP... more ABSTRACT This work describes the practical implementation of two maximum power point tracking (MPPT) algorithms, namely those of perturb and observe, and extremum seeking control. The proprietary dSPACE system is used to perform hardware in the loop (HIL) simulation whereby the two control algorithms are implemented using the MATLAB/Simulink (Mathworks, Natick, MA) software environment in order to control a synchronous buck–boost converter connected to two commercial thermoelectric modules. The process of performing HIL simulation using dSPACE is discussed, and a comparison between experimental and simulated results is highlighted. The experimental results demonstrate the validity of the two MPPT algorithms, and in conclusion the benefits and limitations of real-time implementation of MPPT controllers using dSPACE are discussed.

Research paper thumbnail of Investigation of Maximum Power Point Tracking for Thermoelectric Generators

Journal of Electronic Materials, 2013

ABSTRACT In this paper, a thermoelectric generator (TEG) model is developed as a tool for investi... more ABSTRACT In this paper, a thermoelectric generator (TEG) model is developed as a tool for investigating optimized maximum power point tracking (MPPT) algorithms for TEG systems within automotive exhaust heat energy recovery applications. The model comprises three main subsystems that make up the TEG system: the heat exchanger, thermoelectric material, and power conditioning unit (PCU). In this study, two MPPT algorithms known as the perturb and observe (P&O) algorithm and extremum seeking control (ESC) are investigated. A synchronous buck–boost converter is implemented as the preferred DC–DC converter topology, and together with the MPPT algorithm completes the PCU architecture. The process of developing the subsystems is discussed, and the advantage of using the MPPT controller is demonstrated. The simulation results demonstrate that the ESC algorithm implemented in combination with a synchronous buck–boost converter achieves favorable power outputs for TEG systems. The appropriateness is by virtue of greater responsiveness to changes in the system’s thermal conditions and hence the electrical potential difference generated in comparison with the P&O algorithm. The MATLAB/Simulink environment is used for simulation of the TEG system and comparison of the investigated control strategies.

Research paper thumbnail of A comparative Study of Model-Based and Data-Based Model Order Reduction Techniques for Nonlinear Systems

Advances in Intelligent Systems and Computing, 2014

New technologies and multi-physical description of subsystems have forced designers to consider n... more New technologies and multi-physical description of subsystems have forced designers to consider nonlinear effects for more accurate modelling leading to increased complexity of mathematical models . Such complex models are non-trival to analyse and to develop control algorithms. Consequently, increasing complexity of circuit designs causes the need for model order reduction (MOR) techniques that are capable of reducing nonlinear models and decreasing computational cost of simulating nonlinear systems. MOR techniques for linear time invariant (LTI) systems are well established [2]. On the other hand MOR for nonlinear systems is an open problem [1]. There are several ways of obtaining reduced order model (ROM) for nonlinear systems via model-based approach such as linear approximation(LA) [3], bilinearisation, proper orthogonal decomposition (POD), quadratic approximation (QA) and trajectory piecewise linear (TPWL) approximation, etc.