Prediction using inelastic constitutive artificial neural networks for impact of dead-time on inverters in wireless power transfer systems
Dead time in Wireless Power Transfer Systems (WPTS) affects inverters in several ways, including higher switching losses, electromagnetic interference, and system instability, necessitating innovative mitigation strategies for enhanced efficiency and reliability. Dead time increases switching losses...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-06-01
|
| Series: | e-Prime: Advances in Electrical Engineering, Electronics and Energy |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2772671125001007 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Dead time in Wireless Power Transfer Systems (WPTS) affects inverters in several ways, including higher switching losses, electromagnetic interference, and system instability, necessitating innovative mitigation strategies for enhanced efficiency and reliability. Dead time increases switching losses in inverters, leading to wasted energy and decreased efficiency. This manuscript proposes a novel approach to analyzing the effect of Dead Time on Inverters in WPTS. The proposed approach is the combined operation of both the Snooker-Based Optimization Algorithm (SBOA) and the Inelastic Constitutive Artificial Neural Networks (ICANN). Hence it will be named SBOA-ICANN. This proposed method's main goal is to increase the system's overall efficiency while reducing power losses and system errors. The proposed SBOA is used to optimize the WPTS at its resonant frequency. The ICANN approach enables accurate predictions regarding the impact of dead time on system performance. By then, the MATLAB working platform is used to apply the proposed strategy, and the existing system is used to calculate the execution. The proposed technique displays better results compared to all existing methods such as Feed-Forward Neural Network (FNN), Passive Compensation Network (PCN) and Deep Neural Network (DNN). The existing strategy specify an error of 3.8 %, 4 %, and 4.2 % whereas the proposed strategy specify an error of 3.2 %. The existing strategy specify an efficiency of 90 %, 87 %, and 78 %, and the proposed strategy shows an efficiency of 97 %, which highlighting high efficiency and low error compared to existing methods. |
|---|---|
| ISSN: | 2772-6711 |