PIER C | PIER Journals

2024-12-11

PIER C

Vol. 151, 33-43, 2025

download: 17

Design and Analysis of Linear Primary Permanent Magnet Vernier Machines with Different Winding Configurations

Zhijian Ling , Qi Zhang and Meimei Xu

This paper investigates the effects of winding configurations on force density and fault tolerance in linear primary permanent magnet vernier (LPPMV) machines. Firstly, the LPPMV machines with integral slot distributed windings (ISDWs) and fractional slot concentrated windings (FSCWs) are discussed. Due to the high modulation ratio of ISDW machine, it has the potential to achieve higher thrust force capabilities. Then, the operation principle of the LPPMV machines is analyzed from the perspective of air-gap magnetic modulation. Furthermore, it should be noted that the winding configurations of ISDW machine has larger spans, resulting in insufficient fault-tolerance. To solve this limitation, a new modular ISDW LPPMV machine was proposed and optimized. In the modular ISDW LPPMV machine, a 3×3-phase winding configuration is employed. It is demonstrated that modular ISDW LPPMV machines exhibit superior characteristics in both thrust density and fault tolerance. Finally, the experiments are carried out in a linear test bench, verifying the theoretical analysis.

Design and Analysis of Linear Primary Permanent Magnet Vernier Machines with Different Winding Configurations

2024-12-11

PIER C

Vol. 151, 25-31, 2025

doi:10.2528/PIERC24101302

download: 11

A Novel Proof-of-Concept AI-Driven Approach for Advanced Electromagnetic Imaging

Ali Ghaffarpour , Tahereh Vasei , Mahindra Ganesh , Reza K. Amineh and Maryam Ravan

This paper introduces an artificial intelligence (AI) methodology designed to enhance the output of two-dimensional (2D) electromagnetic imaging systems, specifically tailored for the imaging of conductive objects utilizing inductive sensors. The core of our imaging system comprises a commercial data acquisition board, alongside custom-made multilayer planar coils developed by conventional printed circuit board technology. By leveraging recent advances in AI and machine learning, our approach significantly improves the resolution and clarity of electromagnetic images. The paper uses a multi-layer perceptron (MLP) classifier to process the raw electromagnetic data captured by the imaging system. These algorithms are trained to recognize patterns and anomalies in electromagnetic field data, which are often indicative of conductive objects. The enhanced imaging capability is demonstrated through a series of experiments that compare the AI-enhanced outputs with the ground truth.

A Novel Proof-of-Concept AI-Driven Approach for Advanced Electromagnetic Imaging

2024-12-11

PIER C

Vol. 151, 13-24, 2025

doi:10.2528/PIERC24102605

download: 9

Application of Attention Mechanism-Enhanced BiLSTM-CNN in Power Amplifier Behavioral Modeling and Predistortion

Jingchang Nan , Shize Liu and Jiadong Yu

Power amplifiers in wireless communication systems can introduce nonlinear distortion, degrade signal transmission quality, and increase power consumption. The paper presents a BiLSTM-CNN-based model for modelling power amplifier behaviour to address this issue. The model uses BiLSTM layers to capture temporal information from the signal data and incorporates a multi-head attention mechanism to focus on different temporal features. Additionally, convolutional layers process global features and reduce model parameters through weight sharing. Using this model, a digital pre-distortion (DPD) model is proposed to linearise the power amplifier through an indirect learning approach. The results show that the BiLSTM-CNN model achieves a normalised mean square error (NMSE) of -40.3dB, and the DPD model enhances the adjacent channel power ratio (ACPR) of the communication system by 18dB, demonstrating the model's feasibility. Comparative analysis with other network models indicates that BiLSTM-CNN outperforms traditional methods of fitting performance and convergence speed, showcasing its superiority.

Application of Attention Mechanism-enhanced BiLSTM-CNN in Power Amplifier Behavioral Modeling and Predistortion

2024-12-08

PIER C

Vol. 151, 1-12, 2025

doi:10.2528/PIERC24101603

download: 28

AI-Tuned Metantenna Antenna for Fifth Generation & Beyond Communication Applications

Bikash Ranjan Behera and Harikkrishna Paik

For the purpose of fifth-generation and beyond communication applications, broadband circularly polarized (CP) & high gain AI-tuned metantenna operating in the 5 GHz band is presented in this article. So, an linearly polarized (LP) printed monopole antenna is being taken into consideration in the initial stage. To initiate CP from LP, a metallic strip that functions as a dynamic switching mechanism is utilized to short one of the parasitic conducting strips (PCS) with partial ground plane. The objective is to enhance the impedance (IBW) and axial bandwidths (ARBWs) as well as the antenna gain in order to make it a suitable candidate for ambient RF energy harvesting/wireless energy harvesting application. To achieve this, AI-tuned metasurface is placed below the monopole radiator at a height of 0.33λ_o. With a measured 49.84% IBW, 22.36% ARBW, CP gain > 8 dBic, antenna efficiency > 70%, fabricated on an FR-4 substrate with 1.3λ_o x 0.9λ_o x 0.02λ_o, it is suitable for the technological deployments in a current wireless technology, assuring resilience in networks. To meet the ever-increasing requirements of the current scenario, wireless communication landscape is on a paradigm shift. This transformation is brought by the utilization of metasurfaces offering customized, effective, and typical control of electromagnetic waves keeping with the desired frequency conditions.