14:00
Session 3: Software Implementation
Chair: Kees Vuik
14:00
20 mins
|
PSO ALGORITHMS AND GPGPU TECHNIQUE FOR ELECTROMAGNETIC PROBLEMS
Anton Duca, Laurentiu Duca, Gabriela Ciuprina, Egemen Yilmaz, Tolga Altinoz
Abstract: This paper studies the efficiency of the General Purpose Computation on Graphics Processing Units (GPGPU) technique for the implementation of a parallel Particle Swarm Optimization (PSO) algorithm applied for the optimization of electromagnetic field devices. Several sequential PSO algorithms are compared in order to find the optimal configuration of an electromagnetic device, the TEAM22 benchmark electromagnetic problem. The best PSO algorithm is parallelized by using a GPGPU technique using various configurations for kernels and threads per block. Details of the parallel implementations are explained. The sequential and parallel implementations are compared using as criteria the speed up and the solution quality. The most efficient approach turned to be the one with one thread per block, which was up to 4 times faster than a sequential implementation running on hardware architectures with processors much more advanced than the core processors of the GPU.
|
14:20
20 mins
|
AUTOMATIC THERMAL MODEL GENERATOR FOR MULTIPHYSICS SIZING OPTIMIZATION
Arnaud Baraston, Laurent Gerbaud, Vincent Reinbold, Thomas Boussey, Frédéric Wurtz
Abstract: This paper deals with the multiphysics modeling of mechatronic systems (building energy management, electrical drive). It focuses on the automatic modeling of thermal phenomena using equivalent circuit description. From such a description, the paper proposes an approach able to automatically generate files with the corresponding model equations and their Jacobian (useful for optimization algorithm like SQP). For multiphysics problems, the thermal model is coupled to other models thanks to CADES environment. A main specificity of the approach is the ability to deal with the selectivity of the inputs and outputs of the generated model according to the problem specifications, which reduces drastically the size of the model and the computational complexity. The approach is applied on the sizing of an electrical machine.
|
14:40
20 mins
|
TIME REDUCTION OF THE DYNAMIC PROGRAMMING COMPUTATION IN THE CASE OF HYBRID VEHICLE
Emmanuel Vinot
Abstract: Deterministic Dynamic Programming is frequently used to solve the management problem of hybrid vehicles (choice of mode and power sharing between thermal and electric sources). However, it is time consuming and thus difficult to use in global sizing optimization or in parametric studies. This paper presents a comparison between three methods to compute the DDP problems. These methods are applied on the well known case of the Toyota PRIUS. It proves that a dense matrix method can reduce the computation time by up to 10 compared to classical solving methods.
|
15:00
20 mins
|
OPTIMAL DESIGN OF CORRECTION COILS IN TOKAMAK FUSION DEVICES
Giuseppe Calabrò, Andrea Gaetano Chiariello, Alessandro Formisano, Francesco Ledda, Raffaele Martone, Francesco Pizzo, Giuseppe Ramogida
Abstract: Flux density field map in magnetically confined thermonuclear fusion devices must be carefully designed, in order to achieve the required performance. The ripple phenomenon, due to the modularity of the toroidal field sources, must be reduced using suitably designed “additional” coils. In this paper, the optimization of the shape of such additional coils is tackled, with reference to the FAST Tokamak
|
|
