Virtual vector-based neural network DTC scheme for dynamic performance improvement of dual-star induction motor drive

e-Prime - Advances in Electrical Engineering Electronics and Energy, Год журнала: 2025, Номер unknown, С. 100938 - 100938

Опубликована: Март 1, 2025

Язык: Английский

---

Dynamic behavior and modulation instability of the generalized coupled fractional nonlinear Helmholtz equation with cubic–quintic term

Open Physics, Год журнала: 2025, Номер 23(1)

Опубликована: Янв. 1, 2025

Abstract In this work, the main objective is to study dynamic behavior of generalized coupled fractional nonlinear Helmholtz equation with cubic–quintic term, which describes soliton propagation in optics through theory dynamical system. First, original transformed into an integer-order partial differential using Atangana’s derivative. Then, considering traveling wave and linear transformations, a two-dimensional planar system constructed. Through phase portraits, stability analysis, parameter sensitivity studied. Next, under triangular periodic perturbation logarithmic respectively, reasons for evolution patterns different initial conditions are analyzed. qualitative we have avoided limitations errors precise solution methods, obtained equilibrium point changes, as well system, including chaotic behaviors. Finally, by investigating modulation instability maintaining analysis method.

Язык: Английский

---

Adaptive Morphing Activation Function for Neural Networks

Fractal and Fractional, Год журнала: 2024, Номер 8(8), С. 444 - 444

Опубликована: Июль 29, 2024

A novel morphing activation function is proposed, motivated by the wavelet theory and use of wavelets as functions. Morphing refers to gradual change shape mimic several apparently unrelated The controlled fractional order derivative, which a trainable parameter be optimized in neural network learning process. Given function, taking only integer-order derivatives, efficient piecewise polynomial versions existing functions are obtained. Experiments show that performance PolySigmoid, PolySoftplus, PolyGeLU, PolySwish, PolyMish similar or better than their counterparts Sigmoid, Softplus, GeLU, Swish, Mish. Furthermore, it possible learn best from data optimizing fractional-order derivative with gradient descent algorithms, leading study more general formula based on calculus build adapt properties useful machine learning.

Язык: Английский

---

Analysis of a Solar Photovoltaic System Using Fractional-Order Neural Network

Communications in computer and information science, Год журнала: 2024, Номер unknown, С. 192 - 206

Опубликована: Янв. 1, 2024

Язык: Английский

---