Particle.news
Download on the App Store
4 ARTICLES
·
2mo ago

AI-Driven Design Yields Load-Independent Wireless Power Transfer with High Efficiency

The system’s ability to maintain zero-voltage switching with stable output under varying loads demonstrates a viable path toward commercial wireless power.

image: ©da-kuk | iStock
Researchers developed a fully numerical design method using differential equations and genetic algorithms to optimize WPT systems. This approach ensures stable output voltage, high efficiency, and zero-voltage switching across varying loads, overcoming the limitations of traditional analytical-based methods. Image credit & source link: Wikimedia Commons via Creative Commons Search Repository
Researchers have devised a stable wireless power transfer system

Overview

  • Researchers led by Professor Hiroo Sekiya developed a fully numerical design procedure that models real-world component behavior with differential equations and optimizes circuit parameters using a genetic algorithm.
  • When applied to a class-EF wireless power system, the AI-designed prototype preserved zero-voltage switching and limited output voltage fluctuations to under 5% across load changes.
  • At its rated operating point, the load-independent system delivered over 23 W at 6.78 MHz with an efficiency of 86.7%.
  • Accurate modeling of diode parasitic capacitance and other nonideal effects ensured reliable performance even at light load conditions.
  • The team projects that this machine learning–driven approach could enable cost-effective, scalable wireless power solutions in consumer and industrial devices within five to ten years.