Quasistatic hysteresis modeling with feed-forward neural networks: Influence of the last but one extreme values

Sixdenier, Fabien; Scorretti, Riccardo; Marion, Romain; Morel, Laurent

doi:10.1016/j.jmmm.2008.04.076

Cited by 23 publications

(4 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Additionally, these authors used radial basis function ANNs to model hysteresis in a piezoceramic actuator [8]. In [9], Sixdenier et al used a feed-forward ANN to model scalar hysteresis in a rate-independent manner. Li Y. et al employed a back propagation ANN to combine Jiles-Atherton and Preisach models for hysteresis modelling.…”

Section: Introductionmentioning

confidence: 99%

Model of Shape Memory Alloy Actuator with the Usage of LSTM Neural Network

Rączka,

Sibielak

2024

Materials

View full text Add to dashboard Cite

Shape Memory Alloys (SMAs) are used to design actuators, which are one of the most fascinating applications of SMA. Usually, they are on-off actuators because, in the case of continuous actuators, the nonlinearity of their characteristics is the problem. The main problem, especially in control systems in these actuators, is a hysteretic loop. There are many models of hysteresis, but from a control theory point of view, they are not helpful. This study used an artificial neural network (ANN) to model the SMA actuator hysteresis. The ANN structure and training method are presented in the paper. Data were generated from the Preisach model for training. This approach allowed for quick and controllable data generation, making experiments thoroughly planned and repeatable. The advantage and disadvantage of this approach is the lack of disturbances. The paper’s main goal is to model an SMA actuator. Additionally, it explores whether and how an ANN can describe and model the hysteresis loop. A literature review shows that ANNs are used to model hysteresis, but to a limited extent; this means that the hysteresis loop was modelled with a hysteretic element.

show abstract

Section: Introductionmentioning

confidence: 99%

Model of Shape Memory Alloy Actuator with the Usage of LSTM Neural Network

Rączka,

Sibielak

2024

Materials

View full text Add to dashboard Cite

show abstract

“…In addition to the experimental verification, a computationally efficient vector hysteresis model based on feedforward neural networks (NNs) was utilized to reproduce the behavior of the test sample and to predict power losses. The use of neural networks in simulating hysteresis phenomena has been extensively studied in the literature, mostly for scalar problems [17][18][19][20][21][22][23]. Indeed, NN-based approaches are computationally efficient simulation tools and can be easily formulated in either direct (H input-B output) or inverse (B input-H output) form.…”

Section: Introductionmentioning

confidence: 99%

“…Some of the proposed approaches are actually coupled models, in which the total magnetization is given as the sum of a memoryless component (reversible) computed by the neural network, and an irreversible component determined by exploiting other hysteresis models, such as the Preisach model [17][18][19]. However, other authors used a standalone feedforward neural network to reproduce the hysteresis phenomenon with satisfactory results in one dimension (1-D) [20][21][22]. In most cases the memory storage mechanism is taken into consideration, exploiting feedback algorithms such that the previous value of the model output (or some other information about the "past history") is also given as an input.…”

Section: Introductionmentioning

confidence: 99%

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

et al. 2021

View full text Add to dashboard Cite

A thorough investigation of the 2-D hysteresis processes under arbitrary excitations was carried out for a specimen of innovative Fe-Si magnetic powder material. The vector experimental measurements were first performed via a single disk tester (SDT) apparatus under a controlled magnetic induction field, taking into account circular, elliptic, and scalar processes. The experimental data relative to the circular loops were utilized to identify a vector model of hysteresis based on feedforward neural networks (NNs), having as an input the magnetic induction vector B and as an output the magnetic field vector H. Then the model was validated by the simulation of the other experimental hysteresis processes. The comparison between calculated and measured loops evidenced the capability of the model in both the reconstruction of the magnetic field trajectory and the prediction of the power loss under various excitation waveforms. Finally, the computational efficiency of the model makes it suitable for future application in finite element analysis (FEA).

show abstract

“…To face this problem, a possible alternative is to use "black box" approaches, fully numeric and away from any physical interpretation of the phenomena, but usually fast and reliable for a given numeric reconstruction of data. Among these approaches, we propose in this paper, [9][10][11][12][13][14][15][16][17][18][19][20] a suitably trained Neural System (NS) composed by more than one Feed-Forward Neural Networks (FFNNs) for the modeling of two-dimensional magnetic hysteresis. The NS can be coupled as constitutive functional of the magnetic materials to a FEM solver.…”

mentioning

confidence: 99%

A neural approach for the numerical modeling of two-dimensional magnetic hysteresis

Cardelli

Faba

Laudani

et al. 2015

Journal of Applied Physics

View full text Add to dashboard Cite

This paper deals with a neural network approach to model magnetic hysteresis at macro-magnetic scale. Such approach to the problem seems promising in order to couple the numerical treatment of magnetic hysteresis to FEM numerical solvers of the Maxwell's equations in time domain, as in case of the non-linear dynamic analysis of electrical machines, and other similar devices, making possible a full computer simulation in a reasonable time. The neural system proposed consists of four inputs representing the magnetic field and the magnetic inductions components at each time step and it is trained by 2-d measurements performed on the magnetic material to be modeled. The magnetic induction B is assumed as entry point and the output of the neural system returns the predicted value of the field H at the same time step. A suitable partitioning of the neural system, described in the paper, makes the computing process rather fast. Validations with experimental tests and simulations for non-symmetric and minor loops are presented

show abstract

Quasistatic hysteresis modeling with feed-forward neural networks: Influence of the last but one extreme values

Cited by 23 publications

References 9 publications

Model of Shape Memory Alloy Actuator with the Usage of LSTM Neural Network

Model of Shape Memory Alloy Actuator with the Usage of LSTM Neural Network

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

A neural approach for the numerical modeling of two-dimensional magnetic hysteresis

Contact Info

Product

Resources

About