Implementation of Vector Hysteresis Model Utilizing Enhanced Neural Network Based on Collaborative Algorithm

Abstract

A hysteresis model, based on the enhanced neural network with parallel strategy, is put forward for the prediction of the accurate magnetic behavior of electrical steel sheets (ESSs). Aimed at overcoming the drawbacks such as low convergence rate and convenient to trap into local optimum in the conventional back-propagation neural network (BPNN), a novel collaborative BPNN learning algorithm is introduced according to the error back propagation mechanism and particle swarm optimization (PSO). The reasonable selection of the test point set by the uniform design of experiment methodology, has the potential of lowering the measurement cost, together with guaranteeing the accuracy of the hysteresis modeling. A parallel strategy, which is based on the fast Fourier transformation (FFT), is applied for enhancing the train efficiency of BPNNs. The proposed algorithm is applied for the purpose of modeling the vector hysteresis behavior of ESS. Together, the comparison of the measured and predicted results of H-locus and core loss is discussed as well. INDEX TERMS Back-propagation neural network, collaborative algorithm, design of experiment, parallel strategy, particle swarm optimization, vector hysteresis model.

Comparison between different models of magnetic hysteresis in the solution of the TEAM 32 problem

“…In many cases the non‐linearity of the magnetic core, together with the presence of non‐sinusoidal and strongly distorted current and voltage waveforms becomes a critical issue of the design. Moreover, the real magnetic field in these applications is a vector quantity, and the magnetic hysteresis modeling should be conveniently adapted to that 8–17 . In literature, many studies dealing with this issue have been reported.…”

“…Moreover, the real magnetic field in these applications is a vector quantity, and the magnetic hysteresis modeling should be conveniently adapted to that. [8][9][10][11][12][13][14][15][16][17] In literature, many studies dealing with this issue have been reported. Analytical formulas based on the losses separation criterion have been proposed: the magnetic power losses are separated into static hysteresis losses and dynamic losses.…”

Computing Frequency-Dependent Hysteresis Loops and Dynamic Energy Losses in Soft Magnetic Alloys via Artificial Neural Networks

“…The principal scope of this research is to give a contribution to the development of the dynamic models of hysteresis for innovative soft ferromagnetic materials, exploiting artificial neural networks (ANNs). Until now, ANNs have been successfully applied in the development of both scalar [23][24][25] and vector models of static hysteresis [26][27][28], but fewer studies also take the rate dependence into account [29][30][31][32]. The main advantages of neural network-based models are related to their cheap memory allocation and high computational speed, especially when implemented at a low level of abstraction.…”

Vector Hysteresis Processes for Innovative Fe-Si Magnetic Powder Cores: Experiments and Neural Network Modeling

“…Despite the large amount of papers dealing with the reproduction of the scalar hysteresis phenomenon via NN-based models, the same approaches for vector hysteresis problems have been less explored [25]. The application of a neural system consisting of an assembly of feedforward networks was proposed in [26] to reproduce vector magnetization patterns for Fe-Si laminated steels, but the comparison with the experimental data only covered the rotational loops under circular magnetic induction with sinusoidal waveforms. In addition, the problem of power loss prediction has not been examined.…”

“…The second scope of the work is therefore to deal with hysteresis modeling and power loss estimation, taking into account generic supply conditions. The vector NN model, similar to the one introduced in [26] and also discussed in [28], was implemented in a Matlab ® computing environment and is presented in Section 3. It is seen that the model can be opportunely identified only by exploiting the rotational loops with circular magnetic induction trajectories, whereas the other measured processes can be used as test cases for validation.…”