Alireza Fatemi

Large-Scale Design Optimization of PM Machines Over a Target Operating Cycle

A large-scale finite element model-based design optimization algorithm is developed for improving the drive-cycle efficiency of permanent magnet (PM) synchronous machines with wide operating ranges such as those used in traction propulsion motors. The load operating cycle is efficiently modeled by using a systematic k-means clustering method to identify the operating points representing the high-energy-throughput zones in the torque-speed plane. The machine performance is evaluated over these cyclic representative points using a recently introduced computationally efficient finite element analysis, which is upgraded to include both constant torque and field-weakening operations in the evaluation of the machine performance metrics. In contrast with the common practice, which aims at enhancing the rated performance, the entire range of operation is considered in the present design optimization method. Practical operational constraints imposed by the voltage and current limits of the motor-drive system, excessive PM demagnetization, and torque ripple are accounted for during the optimization process. The convergence to the optimal design solutions is expedited by utilizing a new stochastic optimizer. The developed design algorithm is applicable to any configuration of sinewave-drive PM and synchronous reluctance motors over any conceivable load profile. Its effectiveness is demonstrated by optimizing the well-established benchmark design represented by the Toyota Prius Gen. 2 interior PM motor configuration over a compound operating cycle consisting of common U.S. driving schedules. Multiphysics electromagnetic, thermal, and mechanical performance of the optimized design solutions is discussed in a postdesign optimization stage.

Optimal Design of IPM Motors With Different Cooling Systems and Winding Configurations

Performance improvement of permanent magnet (PM) motors through optimization techniques has been widely investigated in the literature. Oftentimes the practice of design optimization leads to derivation/interpretation of optimal scaling rules of PM motors for a particular loading condition. This paper demonstrates how these derivations vary with respect to the machine ampere loading and ferrous core saturation level. A parallel sensitivity analysis using a second-order response surface methodology followed by a large-scale design optimization based on evolutionary algorithms are pursued in order to establish the variation of the relationships between the main design parameters and the performance characteristics with respect to the ampere loading and magnetic core saturation levels prevalent in the naturally cooled, fan-cooled, and liquid-cooled machines. For this purpose, a finite-element-based platform with a full account of complex geometry, magnetic core nonlinearities, and stator and rotor losses is used. Four main performance metrics including active material cost, power losses, torque ripple, and rotor PM demagnetization are investigated for two generic industrial PM motors with distributed and concentrated windings with subsequent conclusions drawn based on the results.

Fast Multi-Objective CMODE-Type Optimization of PM Machines Using Multicore Desktop Computers

Large-scale design optimization of electric machines is oftentimes practiced to achieve a set of objectives, such as the minimization of cost and power loss, under a set of constraints, such as maximum permissible torque ripple. Accordingly, the design optimization of electric machines can be regarded as a constrained optimization problem (COP). Evolutionary algorithms (EAs) used in the design optimization of electric machines including differential evolution (DE), which has received considerable attention during recent years, are unconstrained optimization methods that need additional mechanisms to handle COPs. In this paper, a new optimization algorithm that features combined multi-objective optimization with differential evolution (CMODE) has been developed and implemented in the design optimization of electric machines. A thorough comparison is conducted between the two counterpart optimization algorithms, CMODE and DE, to demonstrate CMODEs superiority in terms of convergence rate, diversity and high definition of the resulting Pareto fronts, and its more effective constraint handling. More importantly, CMODE requires a lesser number of simultaneous processing units which makes its implementation best suited for state-of-the-art desktop computers reducing the need for high-performance computing systems and associated software licenses.

Design optimization of IPM machines for efficient operation in extended speed range

Practical considerations in the design of Interior Permanent Magnet (IPM) machines are examined to identify the key objectives for large-scale design optimization of such machines over an extended speed range. The goal of the optimization is to ensure efficient operation over the entire operating range under various performance and operational constraints. First, this paper illustrates that when the non-linear and lossy nature of the machine is considered from the efficiency standpoint, congruity of the characteristic versus the rated current cannot be the ideal criterion for constant power operation. Second, a new design optimization algorithm for constant power operation is developed which pursues efficient and reliable performance in the extended speed range under the rated current and maximum voltage constraints. To include the full non-linear nature of the machine and to address the effects of copper and iron losses on the motor behavior, this study is centered on finite element (FE) method for analysis and consequently optimization of a 50 hp, 48-slot, 8-pole IPM with single-layer v-type magnets which is used in the 2004 Toyota Prius. The results can be generalized to other PM motor configurations with sinusoidal back-emf waveforms.

Fast multi-objective CMODE-type optimization of electric machines for multicore desktop computers

Large-scale design optimization of electric machines is oftentimes practiced to achieve a set of objectives, such as the minimization of cost and power loss, under a set of constraints, such as maximum permissible torque ripple. Accordingly, the design optimization of electric machines can be regarded as a constrained optimization problem (COP). Evolutionary algorithms (EA) used in the design optimization of electric machines including the Differential Evolution, which has received considerable attention during recent years, are unconstrained optimization methods that need additional mechanisms to handle COPs. In this paper, a new optimization algorithm that features Combined Multi-objective Optimization with Differential Evolution (CMODE) has been developed and implemented in the design optimization of electric machines. A thorough comparison is conducted between the two counterpart optimization algorithms, CMODE and DE, to demonstrate the CMODEs superiority in terms of convergence rate and constraint handling. More importantly, CMODE requires a less number of simultaneous processing units which makes its implementation best suited for state-of-the-art desktop computers reducing the need for High Performance Computing systems and associated software licenses.

RSM-DE-ANN method for sensitivity analysis of active material cost in PM motors

In this paper, a numerical technique is developed for sensitivity analysis of active material cost (AMC) in PM motors with distributed and fractional slot concentrated windings. A comprehensive analysis is carried out to identify how the optimal design rules and proportions of IPM motors with sintered NdFeB magnets vary with respect to the changes in the commodity prices of permanent magnet material, copper, and steel. The sensitivities of the correlations between the design parameters and the AMC with respect to the commodity price ranges are investigated based on response surface methodology (RSM) and large-scale design optimization practice using differential evolution (DE) optimizer. An innovative application of artificial neural network (ANN)-based design optimization is introduced. Multi-objective minimization of cost and losses is pursued for an overall of 200,000 design candidates in 30 different optimization instances subjected to different cost scenarios according to a systematic design of experiments (DOE) procedure. An interesting finding is that, despite common expectations, the average mass of steel in the optimized designs is more sensitive to changes in the commodity prices than the masses of copper and rotor PMs.

Identification of design rules for interior PM motors with different cooling systems

The conventional scaling rules for the optimal design of electric machines are best suited for naturally cooled machines with stator winding current densities less than 4A/mm2. In this paper, through a comprehensive sensitivity analysis, first, it is demonstrated that the correlations between some geometric variables and the performance metrics of interior permanent magnet (IPM) motors vary significantly with respect to the stator winding current density. For this purpose, three current densities are selected so as to approximately account for naturally cooled, fan-cooled and liquid-cooled machines. Subsequently, a parameterized IPM motor is optimized at these current densities through a large-scale design optimization algorithm by evaluating a total of 20,000 design candidates. The 100 best designs from each group are then identified and extracted to investigate the scaling rules for the optimal design of such IPM motors with different cooling systems. The outcomes of the study are in correspondence with the conventional design principles for naturally cooled machines. Nevertheless, it is illustrated that these rules vary for fan-cooled and liquid-cooled machines owing to the increased ampere loading, and also heavy saturation of the magnetic core in such machines. A configuration of a 50 hp, 48-slot, 8-pole IPM motor with a single-layer v-type magnet is used as the benchmark of this study.

Large-scale electromagnetic design optimization of PM machines over a target operating cycle

A novel automated design algorithm for application-based optimization of permanent magnet (PM) machines is presented in this paper. The proposed algorithm features precise performance evaluation of the potentially heavily saturated machines at high-energy-throughput operating zones using finite element (FE) techniques. First, the energy consumption function associated with the machines operating cycle is efficiently modeled by a number of representative load points using a k-means clustering algorithm. Subsequently, a new approach is developed to assess the performance of the machine at each representative load point with proper control to conform to practical operational constraints imposed by voltage and current limits of the motor-drive system. The developed algorithm is applicable to the optimization of any configuration of PM and synchronous reluctance motors over any conceivable operating cycle. Its effectiveness is demonstrated by optimizing the well-established reference/benchmark design represented by the 2004 Toyota Prius IPM motor configuration over a compound operating cycle consisting of common US driving schedules.

Design optimization of spoke-type PM motors for Formula E racing cars

This paper presents the performance trade-offs in the design optimization of spoke-type permanent magnet (PM) motors for high speed and very high torque density traction motors. An example 18-slot 16-pole machine for a direct drive Formula E race car over the Le Mans driving cycle is considered. Both low speed and extended speed/field-weakening operations are evaluated using high fidelity finite element (FE) simulations, to simultaneously increase the torque density and decrease the power losses over the high energy-throughput-zones of the machine torque-speed plane. The results of the design optimization process yielding 3,400 design candidates are utilized to quantify the performance trade-offs for increasing the power density in spoke-type PM motors. These trade-offs include the impacts on other performance metrics such as power losses, PM demagnetization, and torque ripple. The analysis is supplemented by multi-physics simulation of three counterpart optimized designs, and successful experimental verification of a prototype of one of those three designs which represents a record high power density motor in traction applications.

A computationally efficient method for calculation of strand eddy current losses in electric machines

In this paper, a fast finite element (FE)-based method for the calculation of eddy current losses in the stator windings of randomly wound electric machines with a focus on fractional slot concentrated winding (FSCW) permanent magnet (PM) machines will be presented. The method is particularly suitable for implementation in large-scale design optimization algorithms where a qualitative characterization of such losses at higher speeds is most beneficial for identification of the design solutions which exhibit the lowest overall losses including the ac losses in the stator windings. Unlike the common practice of assuming a constant slot fill factor, sf, for all the design variations, the maximum sf in the developed method is determined based on the individual slot structure/dimensions and strand wire specifications. Furthermore, in lieu of detailed modeling of the conductor strands in the initial FE model, which significantly adds to the complexity of the problem, an alternative rectangular coil modeling subject to a subsequent flux mapping technique for determination of the impinging flux on each individual strand is pursued. The research focus of the paper is placed on development of a computationally efficient technique for the ac winding loss derivation applicable in design-optimization, where both the electromagnetic and thermal machine behavior are accounted for. The analysis is supplemented with an investigation on the influence of the electrical loading on ac winging loss effects for a particular machine design, a subject which has received less attention in the literature. Experimental ac loss measurements on a 12-slot 10-pole stator assembly will be discussed to verify the existing trends in the simulation results.