10:00
Poster Session
NUMERICAL ANALYSIS ON THE FORCE IN AN LINEAR RELUCTANCE VARIABLE MOTOR
Nolvi Baggio Filho, Tamara Baggio, Roniele Belusso
Abstract: Currently, many industrial processes require a precise linear motion. Usually this movement is achieved with the use of rotary motors combined with an electrical control systems and mechanical systems such as gears, pulleys and bearings. Other types of devices are based on linear motors, where the linear motion is obtained directly. The Linear Stepper Motor (MLP) is an excellent solution for industrial applications that require precise positioning and high speed. This study presents an MLP formed of a linear structure and static ferromagnetic material, and a mover structure which are mounted three coils. Mechanical suspension systems allow a linear movement between static and mover part, maintaining a constant air gap. The operating principle is based on the tendency of alignment of magnetic flux through the path of least reluctance; the force proportional to the intensity of the electric current and the speed proportional to the frequency of the excitation coils. The study of this device is still based on the use of a numerical analysis to verify the relationship between electric current applied and planar and normal force developed. In addition, the magnetic field in the air gap region is also monitored.
|
COMPARISON BETWEEN SYSTEM DESIGN OPTIMIZATION STRATEGIES FOR MORE ELECTRIC AIRCRAFT NETWORKS
Djamel Hadbi, Nicolas Retière, Frédéric Wurtz, Xavier Roboboam, Bruno Sareni
Abstract: Nowadays, embedded aircraft system contains electrical devices which must cooperate in safe and light weight network. For designing such systems, different local strategies have been developed but no global optimization has been performed so far. In this paper, we present and compare three strategies applied to the sizing of a whole network of more electric aircraft: a simplified case study with only two components is considered to illustrate methodological issues. The quality of the solution found from each method is compared, with regards to the “cost of the collaborative approach” and the volume of data generation. This comparison should provide system designers an evaluation of the applicability of these methods according to the nature of the design problem
|
A MULTILEVEL DOUBLE LOOP APPROACH FOR THE DESIGN OF ONBOARD FLIGHT NETWORKS
Djamel Hadbi, Nicolas Retière, Frédéric Wurtz, Xavier Roboam, Bruno Sareni
Abstract: After testing different existing design methods for complex problems, we have concluded that a good approach based on system decomposition must coordinate the design process of components to reach the system optimum. In this paper, we present a multilevel collaborative approach for designing complex systems based on several loops (here 2). A system level optimization loop added to lead optimizations of components at their optimal solutions. This method was applied to the sizing of a simplified embedded electric network with single source–load configuration.
|
MULTI-OBJECTIVE ROBUST OPTIMIZATION OF LOW FREQUENCY ELECTROMAGNETIC DESIGN PROBLEMS USING NEW MULTI-OBJECTIVE ROBUSTNESS MEASURE
Armin Salimi, David Lowther
Abstract: Optimizing for multiple objectives and obtaining a robust design are two challenging and important areas in electromagnetic design optimization and design optimization in general, yet the combination of these two which means achieving a robust multi-objectively optimized design is a subject rarely visited. In this paper a new metric for measuring robustness in a multi-objective space is introduced. The metric is then used as a constraint while performing multi-objective optimization on test problems and the results are presented.
|
CONTINUOUS FLOCK-OF-STARLINGS OPTIMIZATION FOR A GENERAL MAGNETIC HYSTERESIS MODEL
Ermanno Cardelli, Antonio Faba, Antonino Laudani, Francesco Riganti Fulginei, Alessandro Salvini
Abstract: A new general typology of optimization algorithm inspired to the classical swarm intelligence, the Continuous Flock-Of-Starlings Optimization (CFSO), is used to face the inverse problem of modeling magnetic materials. It is obtained by translating the numerical swarm/flock-based algorithms into differential equations in the time domain and employing analytical closed-forms written in the continuum. The modeling of magnetic hysteresis is very important for the development of magnetic materials and the accurate design of electrical machines. Recently several hysteresis vector models have been proposed and discussed; one of them is the Vector Hysteron Model (VHM) which is a natural extension of the Preisach theory to the vector case. In this work we present a CFSO based approach for the identification of VHM in order to reproduce rotational hysteresis loops. Experimental validations are made for both NOG (Not oriented grain) and OG (Oriented grain) steels.
|
INVERSE-PROBLEM BASED PARAMETER ESTIMATION FOR PERMANENT-MAGNET SYNCHRONOUS MACHINES IN DIFFERENT LOADING CONDITIONS
Paavo Rasilo, Ahmed Abdallh, Ahmed Hemeida, Peter Sergeant, Marko Hinkkanen, Luc Dupré
Abstract: The parameter identification for the direct-quadrature (d-q) axis model of a permanent-magnet (PM) synchronous motor is formulated as an electromagnetic inverse problem. The d-q model parameters are solved by comparison of the d-q model to a state-space model utilizing lookup tables of finite-element simulation results. The d-q inductances and the PM flux linkage are estimated for different operating points. The results compare reasonably to measurements.
|
DEMAGNETIZATION CURRENT EVALUATIONS USING FINITE ELEMENT METHOD AND MAGNETIC EQUIVALENT CIRCUIT MODELING AND OPTIMUM DESIGN IN A POLE CHANGING MEMORY MOTOR
Jung Ho Lee, Su Yong Kim, Young Hyun Kim
Abstract: This paper deals with the demagnetization current evaluations using Finite Element Method (FEM) and magnetic equivalent circuit modeling in a Pole Changing Memory Motor (PCMM). The focus of this paper is the characteristics evaluation relative to the current density and the magnet numbers of machine for pole changing condition in PCMMs. Magnetic equivalent circuit modeling method and demagnetizing currents characteristics of PCMM according to the magnet thickness variation are introduced. Finally, this paper deals with optimum design criteria to minimize the torque ripple, to maximize torque density of a PCMM using Response Surface Methodology (RSM).
|
OPTIMUM DESIGN CRITERIA OF VARIABLE FLUX MEMORY MOTOR USING MAGNETIC EQUIVALENT CIRCUIT MODELING AND RESPONSE SURFACE METHODOLOGY
Jung Ho Lee, Jung Woo Kim, Young Hyun Kim
Abstract: This paper deals with the optimum design criteria of a Variable Flux Memory Motor (VFMM) using magnetic equivalent circuit modeling and Response Surface Methodology (RSM). The focus of this paper is found firstly a design solution through the comparison of torque and efficiency according to magnetizing direction and quantity of permanent magnet and dimensions variations and, secondly, a mixed resolution with Central Composite Design (CCD) is introduced and analysis of variance (ANOVA) is conducted to determine the significance of the fitted regression model.
|
COMPARISON OF OPTIMIZATION FORMULATIONS TO DESIGN AN HYBRID RAILWAY POWER SUBSTATION
Maxime Ployard, Stéphane Brisset, Florent Delhaye, Julien Pouget
Abstract: This paper presents the design of a Hybrid Railway Power Substation. Two optimization formulations are investigated to obtain the optimal design and the energy management strategy that minimize the total cost of HRPS during an operating period of one year.
|
INCREASED ENERGY EFFICIENCY THROUGH SUPERVISORY CONTROL OF STORAGE COMPONENTS MODELLED USING HYBRID PETRI NETS
Guillaume Crevecoeur, Annelies Coene, Luc Dupré
Abstract: To accommodate the fluctuating demand with variable energy sources, means of energy storage need to be added into the energy grid. The complexity of the interactions between the various components in an energy grid need to be efficiently modeled for simulation purposes and for increasing the energy efficacy of power grids. In order to take into account the hybrid nature of the system, i.e. discrete and continuous time-varying processes, the modeling is performed using Petri nets adapted for energy grids. Supervisory control is implemented to attain optimal energy dynamics. This is implemented by finding the optimal control places and transitions in Petri net formulations that alter the connectivity between the various components and thus the topology of the system. We were able to reduce the use of conventional energy production by 5.5% using the supervisory controller.
|
NUMERICAL PERFORMANCE OF HOLE SENSITIVITY-BASED LEVEL SET METHOD FOR TOPOLOGY OPTIMIZATION IN ELECTROMAGNETIC SYSTEM
Seung Geon Hong, Kang Hyouk Lee, Il Han Park
Abstract: While the continuum shape sensitivity with the level set method for topology design optimization has been used in electromagnetic systems, the level set method coupled with the finite element method has a limitation, whereby the degree of freedom decreases gradually during an iterative design process. In this paper, we present the usefulness of hole sensitivity, which was derived in a recent paper, in electromagnetic systems. To show the usefulness of hole sensitivity, the level set method with hole sensitivity is applied to the shape and topology optimization of a transformer and synchronous reluctance motor. The numerical results of the design model are compared with those without hole sensitivity. The results show that the performance of optimization with hole sensitivity is better than that without.
|
RECONSTRUCTION OF 3D DIELECTRIC OBJECTS BURIED UNDER 2D ROUGH SURFACES BY USING CONTRAST SOURCE INVERSION METHOD
Yasemin Altuncu
Abstract: A numerical method based on contrast source inversion (CSI) algorithm is proposed to reconstruct complex dielectric permittivity variation of an inaccessible 3D dielectric object buried under 2D rough surface. The initial step of the method is to introduce the data and object equations involving the dyadic Green’s function of the background medium which consist of a rough interfaced half-space. The Green’s function is calculated numerically by buried object approach (BOA). Then data and object equations are solved iteratively as updating unknown contrast source and contrast function at each step for minimization of a defined cost functional. Some numerical implementations are given to exhibit the performance of the presented method.
|
POWER FLOW OPTIMIZATION IN A MICRO-GRID WITH TWO KINDS OF ENERGY STORAGE
Remy Rigo-Marianni, Vincenzo Roccuzzo, Bruno Sareni, Maurizio Repetto, Xavier Roboam
Abstract: The paper presents the optimization of the power flows inside a micro-grid with renewable sources and two kinds of storage. The optimization is carried out on a period of one day at one hour discretization and the variables are the degrees of freedom that allow controlling the whole system in order to minimize the electrical bill. Different optimization approaches are used and results compared both in terms of accuracy and of computational efficiency.
|
SELECTIVITY AND IMPROVEMENT OF THE MODELLING OF POWER ELECTRONICQ FREQUENCY MODELS FOR OPTIMISATION
Laurent Gerbaud, Jean-Luc Schanen
Abstract: The paper deals with frequency modelling of power electronics for the sizing by optimization. Generally, a model is not dedicated to particular specifications and optimization algorithms. If the specifications for the optimization are not considered at the modelling step, this may induce great useless modelling and computation. This may also limit the sizing great applications, due to large memory consumption and large CPU time. The paper proposes to formulate the sizing model, not in a general way, but by considering the optimization algorithm used and the specifications of the sizing at the modelling steps. So, there are the inputs and outputs for the generated model. The paper focuses on the automatic building of EMC frequency models and applied it to the design of electrical systems of aircraft.
|
AID FOR THE MODELING OF AC SINGLE-PHASE POWER GRID FOR STUDIES USING OPTIMISATION
Hoa Xuan Nguyen, Abir Rezgui, Laurent Gerbaud, Nicolas Retière, Frédéric Wurtz
Abstract: Purpose This paper deals with an automatic modelling of power grid for studies using optimization algorithms for optimal power flow, design, or both. The optimization of such an application is often carried out using Interior Point method, due to the non-linearity of the modelling and the large number of constraints (some hundreds of thousands). In the paper, two implementations of such a method are used: Knitro (Nocedal et al, 2006) (from GAMS (GAMS)) and IPOpt (Wächter et al, 2006). The modelling is based on a nodal approach.
Design/methodology/approach The models are generated in two formats: the modelling language of GAMS and under java native code for the optimization with IPOpt. The gradient of the model is also automatically built in the Java model. The approach has been applied on a power grid reference defined in the literature
Findings A main specificity of the proposed approach is its ability to deal with the building of the equations and their gradients, with a model oriented for optimal power flow and included in global sizing of electrical power grids.
Research limitations/implications At the present time, for the first modelling prototype, only alternative single-phase power grids are considered. The model Hessian is approximated by IPOpt, but should be obtained in an exact way.
Originality/value The modelling process from a power grid schematic is optimization-oriented.
Keywords: Power Grid, Modelling, OPF, Interior point optimization
|
Optimal Power Flow with Storage in DC Electrified Railways
Olivier Bossi, Julien Pouget, Nicolas Retière, Laurent Gerbaud
Abstract: Purpose: Due to the increase of the traffic, issues are appearing on DC electrified railway feeding systems. One candidate solution for solving these issues and improving their performances is the addition of storage systems in the railway DC electrical network. This paper presents a method based on an Optimal Power Flow (OPF) for analysis and design of DC railway feeding systems with storage.
Design / methodology / approach: This paper describes a new methodology to study DC electrified railways including storage systems, based on optimization. A load flow model of a DC 1500V railway electrification system is presented, including the mobility of the train sets. Then, an Optimal Power Flow model of the DC network, including energy storage systems and feeding substation rectifiers has been developed. Finally, the OPF model has been tested on a real application case showing his benefits while searching solutions in order to improve the network performances. Findings: An OPF model suitable for analysis of DC networks with storage is presented. It shows its ability to solve large scale problems. Research limitations/implications: This paper focuses on the physical model of the network. The optimization model will have to be extended with applicative constraints. Originality/value: The hypothesis presented in this paper allows removing the discontinuities of the system in order to use a continuous optimization approach.
|
MULTI-OBJECTIVE OPTIMIZATION IN THE LORENTZ FORCE VELOCIMETRY FRAMEWORK
Dzulia Terzijska, Margherita Porcelli, Gabriele Eichfelder
Abstract: Abstract. We consider an academic problem within the framework of the Lorentz force velocimetry measurement technique. The goal is to increase the Lorentz force in relation to the weight of the magnet system and to determine thereby novel magnet designs. To address this issue we define two multi-objective optimization formulations: one is with continuous variables and the other one is with mixed-integer variables. We investigate and compare these two formulations using the derivative-free NOMAD algorithm.
|
IDENTIFICATION OF PARAMETERS OF A STIFF TURBOGENERATOR MODEL FROM THE NUMERICAL FREQUENCY RESPONCE TEST
Alexei Adalev, Nikolay Korovkin, Sergey Ionin, Mikhail Roitgarts, Andrei Smirnov
Abstract: The work is related to stiff inverse problems of electrical engineering. A well-known problem of identification of parameters of a turbogenerator lumped-circuit model by means of processing the data of Numerical (FEM) Frequency Response test is considered here. It has been shown that increasing model adequacy (with respect to a better reproducing of the numerical data), leads to rising model stiffness that reveals itself through a ravine-shaped minimized residual functional. The stiff minimization problem is proposed to be solved by means of linear relations between the model parameters allowing to reduce model stiffness and obtain the solution with a reliable tolerance.
|
CONVERGENCE OF APPROXIMATE LINEAR SOLVERS FOR ADAPTIVE RUNGE-KUTTA METHODS APPLIED TO THE SIMULATION OF TRANSIENTS IN POWER SYSTEMS
Romain Thomas, Domenico Lahaye, Kees Vuik, Lou van der Sluis
Abstract: Transients in power systems create high frequency oscillations. Modelling to these high frequency oscillations requires small time steps are needed to be applied. These time steps have in term an effect on the conditioning of the linear system to be solved within an adaptive Runge-Kutta method. This motivates to study the speed of convergence of approximate solvers with re-ordering method(AMD). This study suggests that approximate solvers are adequate to solve the linear equations of adaptive Runge-Kutta methods.
|
SENSITIVITY ANALYSIS AND OPTIMIZATION OF PVDF PIEZOELECTRIC MATERIALS WITH INTERDIGITATED ELECTRODES
Nicolas Galopin, Nicolas Choulet, Benoit Delinchant
Abstract: The paper is about efficient sensitivity and optimization procedure applied to piezoelectric device with interdigitated electrodes. Model behavior analysis tools based on Jacobian are more specifically used, and then we are focusing on symbolic Jacobian computation. The electical field modeling is based on conformal mapping which makes appearing semi-analytical expressions based on elliptic integrals and Jacobi special functions.
|
ANALYTICAL FREQUENCY MODEL OF A DIODE RECTIFIER: SQP SOLVING VERSUS NEWTON-RAPHSON SOLVING
Le Nhat Hoang Tran, Laurent Gerbaud, Retiere Nicolas, Nguyen-Huu Hieu
Abstract: Static converters generate current harmonics in grids. Numerous studies on analytical frequency model are preferred to carry out their harmonic modelling in the context of sizing by optimization. Some solving methods are proposed to solve such models. Each solving method has its own advantages and drawbacks. The paper mainly focuses on two approaches: the first with Sequential Quadratic Programming solving (SQP) and the second with Newton-Raphson solving (NR). In this way, the paper presents the performances of each method and compares results of the two methods for the modelling of a single-phase diode rectifier.
|
MULTI-OBJECTIVE OPTIMIZATION OF THE SIZING OF A HYBRID ELECTRICAL VEHICLE
Vincent Reinbold, Laurent Gerbaud, Emmanuel Vinot
Abstract: Hybrid electrical vehicles involve two sources of energy, usually gasoline and electricity. The energy management determines the power sharing between the internal combustion engine (ICE) and the electrical machine (EM). It is highly dependent on the driving cycle (i.e. the use of the vehicle). In this context, the optimal sizing of the EM is determined by: the driving cycle, the power-train characteristics (i.e. ratios and physical limitations e.g. maximum torque available) and the energy management. The key idea of this work is to involve the driving cycle and the environment of the electrical machine in a global multi-objective optimization process taking into account an optimal energy management (OEM).
|
JOINT OPTIMIZATION OF CONTROL AND SIZING OF THE PARALLEL HEV USING SQP ALGORITHM
Vincent Reinbold, Laurent Gerbaud, Emmanuel Vinot
Abstract: In electromagnetic research, the optimization of electrical machine is a very common issue. The improvement of computational capabilities allows researchers to improve models and complexity of the optimization problem. In transportation applications, electrical machines are subjected to a large panel of operating points. Moreover, for hybrid vehicles, the sizing problem involves a control problem as a sub-optimization problem (power sharing between electric and thermal transmission). Therefore, a joint optimization of the energy management and sizing of the EM over a driving cycle shall be considered. The paper presents the modelling of the electrical machine based on magnetic circuit equivalence.
|
DESIGN OPTIMIZATION OF ELECTROMAGNETIC DEVICES USING THE LEAGUE CHAMPIONSHIP ALGORITHM
Houssem Bouchekara, Ahmed Abdallh, Luc Dupré, Hamza Kherrab, Rabia Mehasni
Abstract: Nowadays, there is an increasing attention to novel evolutionary techniques. A new efficient optimization method, called the League Championship Algorithm (LCA) is proposed in this paper for the optimal design of electromagnetic devices. This method is inspired by the competition of sport teams in an artificial sport league for several weeks and over a number of seasons. The validity of the algorithm is tested on the magnetizer benchmark problem, and results are compared with those found in the recent literature, showing the effectiveness and robustness of the proposed method.
|
Optimization of the control of a doubly fed induction machine
Jules Gillet, Maria Pietrzak–David, Frédéric Messine
Abstract: This paper focuses on finding an optimal control for a doubly fed induction machine (DFIM) in a motor mode. Our purpose is to develop a numerical method to improve the DFIM efficiency. Thus, we will start with a fixed system made by a DFIM and two ideal 3 phases voltage inverters. Our method will demonstrate its efficiency to improve that control for minimizing the copper loss. This method is based on direct shooting techniques associated with a MATLAB optimization solver: fmincon.
|
PARAMETRIC OPTIMIZATION OF MAGNETIC CIRCUITS OF HALL EFFECT THRUSTERS
Alberto Rossi, Carole Henaux, Frédéric Messine
Abstract: This study presents a design method based on parametric optimization. Knowing that
the current average of 80 % of the weight of Hall Effect thrusters is the weight of the magnetic circuit, the two fundamental constraints of dimensioning are respectively the minimization of the mass and the obtaining a specific and given magnetic mapping. The design of Hall Effect thrusters until now was based on empirical models. In this paper, we present some theoretical models and mathematic formulations to find optimal magnetic circuits for those thrusters.
|
COMPARATIVE ANALYSIS ON VARIABLE SYNRM ACCORDING TO GEOMETRIC STRUCTURE BY NUMERICAL ANALYSIS AND EXPERIMENTATION
Jung Ho Lee, Jun Ho Lee, Young Hyun Kim
Abstract: This paper presents the loss and efficiency evaluations for synchronous reluctance motors (SynRM) with permanent magnet assisted (PMA), non-PMA and concentrated winding (CW), distributed winding (DW) by finite element method (FEM) and experiment. In this paper, the performance characteristic is compared according to the stator and rotor types in SynRM. In addition, the efficiency evaluation is also compared including hysteresis loss, copper loss, other loss on the basis of rated load condition.
|
OPTIMUM DESIGN CRITERIA OF ALA ROTOR SYNCHRONOUS RELUCTANCE MOTOR FOR THE MAXIMUM TORQUE DENSITY AND POWER FACTOR IMPROVEMENT
Jung Ho Lee, Jin Kyoung Lee, Young Hyun Kim
Abstract: This paper deals with the characteristics analysis and optimum design of axially laminated
anisotropic (ALA) rotor Synchronous Reluctance Motor (SynRM) considering centrifugal force and the
effect on flux path. The rotor core of an ALA type adheres on the shaft axis. Because a rotor rotates at the
rated speed as 3000rpm, the rotor is received with the centrifugal force, about 147.16[kN] per one
block(diameter of the out rotor : 326.6mm). Because it is dangerous to fix the rotor, which is affected
with the centrifugal force into shaft axis, this paper deals with the optimum design of the ALA-type rotor
to solve the problem. Therefore, by simulating the structure and electromagnetic field, it is confirmed that
the width of rotor core and insulator is set to design variables to reduce the minimum loss and the
centrifugal force. In addition, Comparisons are given with characteristics of a same rated wattage traction
induction motor and those of ALA-SynRM, respectively.
|
OPTIMUM DESIGN OF PMA-SynRM ACCORDING TO THE RATED WATTAGE FOR POWER IMPROVEMENT
Jung Ho Lee, Jun Seo, Young Hyun Kim
Abstract: This paper deals with the optimum design criteria of Permanent Magnet Assisted Synchronous Reluctance Motor (PMA-SynRM) for power improvement. The focus of this paper is found a design solution through the comparison of torque density and d- and q-axes inductances according to the rotor magnet, the number of barriers and dimension variations in various rated wattage.
|
NONLINEAR MULTIOBJECTIVE TOPOLOGY OPTIMIZATION AND MULTIPHYSICS ANALYSIS OF A PERMANENT-MAGNET EXCITED SYNCHRONOUS MACHINE
Piotr Putek, Piotr Paplicki, Roland Pulch, Jan ter Maten, Michael Günther, Ryszard Pałka
Abstract: This paper proposes the multiobjective topology optimization of the Electric Controlled Permanent-Magnet Synchronous Machine (ECPMSM) using the level set method and Continuum Design Sensitivity Analysis (CDSA) in order to reduce both the electromagnetic losses and the Cogging Torque (CT). The back-electromotive force (the back EMF) is taken into account in the multiobjective optimization. First, the losses of the ECPMSM such as the stator iron losses, rotor eddy current losses and the hysteresis losses are calculated by the 2D time-harmonic Finite Element (FE) analysis. Next, the temperature distribution is estimated by a coupled heat transfer analysis. During an iterative shape and topology optimization process, the field velocity in the level set method is evaluated using the CDSA approach. Simulation outcomes show that the applied method leads to a significant reduction of both the electromagnetic losses and the CT and additionally minimizes the higher harmonics in the back-EMF.
|
|
