Ignace Rasoanarivo

Design of fractional order PID controller for boost converter based on Multi-Objective optimization

This paper proposes the theory and application of a new designing method of the Fractional Order PID (FOPID) controller for boost converters. FOPID is a PID where the derivation and integration orders are of fractional order rather than integer. A FOPID is an extension to classic integer order PID controllers, that potentially promises better results. The method is based on the use of a Multi-Objective optimization evolutionary algorithm called Strength Pareto Evolutionary Algorithm (SPEA). Other controller methods cannot guarantee a good start-up response but the proposed method can provide an excellent start-up response besides desired dynamic response. The optimum fractional order PID coefficients for the desired control objectives are included and designers can implement each of them based on objective functions priority. A comparison between the optimum integer order PID controller and optimum fractional order PID controller is presented in the paper. The simulation and some experimental results prove the superiority of the fractional controllers over the integer controllers.

Time-Stepping Simulation of Synchronous Reluctance Motors Using a Nonlinear Reluctance Network Method

We present a nonlinear reluctance network approach for the computation of the electromotive force (EMF) waveforms of synchronous reluctance motors (SynRM) with a massive rotor or a flux barrier rotor. We model all ferromagnetic parts of the machine by nonlinear reluctances in order to take the saturation into account. The model of the motor consists of three reluctance networks: of the stator, of the rotor, and of the air gap. The originality of the work lies in the automatic computation of the topology and of the reluctance values of the reluctance network. The computation models the air gap for any relative position of the rotor and the stator; thus, the movement of the rotor can be taken into account. For any saturation level, a comparison with time-stepping finite-element results shows good agreement for the EMF fundamental, the mean torque, and the EMF and torque harmonics of order lower than the slotting ones. For a normal saturation level, the reluctance network models also accurately compute the slotting harmonics.

A novel multilevel inverter model

This paper discuss on a novel multi-level inverter neutral point (MNP) without the diode-clamped which deduced from neutral point clamped (NPC) or multipoint clamped (MPC) topologies. One extremity of the phase branch is linked to different voltage levels of the dc bank capacitor supply by one and only one power switch. In this inverter model, we increase the security and the reliability of the power switches. The first advantage in this model is to remove the diodes clamped per phase that in the classical model we used always the diode-clamped. The second originality is a simple command strategy for this device with a low switching frequency in the purpose of providing feeble commutation and conduction looses and the good output voltage qualities. In addition this inverter can work with different voltage waveforms from two, three until the N-level of the dc supply source. At the end we checked our system with simulations and experimental results.

Reluctance Network Analysis of High Power Synchronous Reluctance Motor with Saturation and Iron Losses Considerations

This paper provides a reluctance network modeling for the steady state performance calculation of a 94kW massive or a 125kW flux barrier rotor synchronous reluctance machine. The massive rotor is suitable for high-speed application; the flux barriers increase the saliency ratio and lead to higher torque and power factor. The suggested model is significantly faster than finite element method and is precise as for performance criteria such as torque and power factor. This precision is guaranteed by taking into account the saturation of all ferromagnetic parts of the machine. The proposed model also permits to compute the stator iron losses and consequently the motor efficiency.

Time Stepping Simulation of Synchronous Reluctance Motors Using Non Linear Reluctance Network Method

This paper provides a non linear reluctance network approach for the computation of the EMF waveforms of synchronous reluctance motors (SynRM) with a massive and a flux barrier rotors. All ferromagnetic parts of the machine are modeled by non linear reluctances in order to take into account the saturation. The model of the motor consists of three parts: the reluctance network of the stator, the reluctance network of the rotor and the reluctance network of the airgap. The originality of the work is the automatic computation of the topology and the reluctance values of the reluctance network which models the airgap for any relative position of the rotor and the stator: thus, the movement of the rotor can be taken into account. Comparison with time stepping finite element results shows a good accordance as for the EMF fundamental and the EMF harmonics of order lower than the slotting ones for any saturation level. For normal saturation level, the reluctance network models compute accurately the slotting harmonics

A new multilevel inverter model NP without clamping diodes

This paper deals with a novel multilevel inverter neutral point (MNP). This system consists of a series of monodirectional and bidirectional switch. If one or several bidirectional power switches of this inverter burn out or to become ruined by reason an over voltage or depreciation, it stays in the previous level. Indeed, for a three-level NP to occur a defect in bidirectional switch (middle switch), this system operate in two level mode before detecting defect. In this inverter model, we increase the security and the reliability of the power switches. The originality of this inverter is to remove clamping diodes per phase in comparison with the classical multilevel inverter model neutral point clamped (NPC): e.g. for three phase of an inverter with n-level we removed 6.(n-2) diode-clamp. The advantage of delete all clamping diodes is to decrease the conduction losses. The other originality of this work is a simple command strategy with a low switching frequency. At the end we compare simulated and experimental results.

Constrained Optimization of High Power Synchronous Reluctance Motor Using Non Linear Reluctance Network Modeling

This paper provides a performance optimization of synchronous reluctance motors (SynRM) up to 130 kW using a non linear reluctance network modeling and a sequential quadratic programming (SQP) minimization technique. The non linear reluctance network approach allows to take into account the saturation and is significantly faster than the finite element method. 3000rpm-2pole machines with either a massive rotor or flux barriers are optimized. The stator and rotor geometrical parameters and the number of turns of the stator windings are simultaneously optimized. A separate torque ripple study using finite element method is carried out. The performance of the optimized motors are finally verified by a time-stepping finite element calculation

Universal Control Strategy for Three Different Multilevel Inverters

The present paper deals with a generalized control strategy for the following multilevel inverters: Neutral Point Clamped (NPC), H-Bridge, and Load Phase Clamped (LPC). The contribution of this work is to develop, in spite of these devices differences, a common control strategy using simple triangular sinus PWM. By applying Petri Network theory, common configurations of these inverters are established. The generation of the trigger commands requires an accurate load current estimation to achieve a safe operating. Results by simulation and by experimentation are shown and discussed to validate the efficiency of the proposed technique.

Strength Pareto Evolutionary algorithm for sizing a set of SiC converter connected to AC machines winding by short planar cable

This paper deals with a set of ultra fast converter connected to the winding of AC machine by a short planar cable. The main feature of this study is to design a set of bus bar, a planar cable to mitigate surges produced at each switching commutation of the SiC Transistor, without consequently modifying its rise time. This calculation is obtained by the principle of genetic algorithm based on Strength Pareto Evolutionary Algorithm (SPEA), with as extra constraint a limited volume of the overall size of the system. Simulations analysis and experimental tests are carried out and compared, they underline the main benefits of the developed method.

High frequency modeling of the winding wires of AC machines

The present paper deals with the modeling in high frequency of AC machines windings, when the coils are subjected to fast slopes of MLI voltage. The purpose of this study is twofold: firstly, to develop a network of the electrical circuit, and secondly, to effectuate a behavioral analysis. The different elements of this network will be identified by Finite Element Calculations thanks to COMSOL MULTIPHYSIC software, and a behavioral analysis of this network is simulated by MATLAB SIMULINK. This analysis confirms the occurrence of voltage spikes in the first few turns of the coil. Finally, an improvement to face these surges is proposed, consisting in modification of the coupling of partial coils of the winding. Simulated and experimental results are matching well and highlight the merits of the present approach.