Scholarlink

<p>The artificial intelligence-based frameworks available to date have more or less failed to provide mechanical fidelity, resilience in real-world operating conditions, and multi-objective optimization capability. A holistic AI-based multi-criteria optimization framework developed in this work consists of Quantum Informed Graph Transformer Networks (QGTN), Self-Supervised Contrastive Multi-Objective Learning (SCMOL), Transformer-Based Evolutionary Surrogate Models (TESM), Physics Informed Generative Adversarial Networks (PI-GANs), and an AI-Driven Cyber-Resilient Digital Twin (AICR-DT). The proposed system augments the energy density of the baseline materials by 49.4%, while also improving cycle lifespan by 86.2%. The prediction accuracy of stability has improved to 94.7%, with a decrease of 97.6% in the generation of unstable structures. Moreover, the screening time by high throughput has been decreased by 85.9%, while real-time adaptive optimization has multiplied the supercapacitor life span by 3.1×. All these results reaffirm the robustness, scalability, and practicality of the proposed framework in accelerating the discovery and deployment of next-generation nanomaterials for high-performance electric vehicle (EV) supercapacitors.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An integrated artificial intelligence-driven approach to multi-criteria optimization of nano-materials for high-capacity electric vehicles supercapacitors

  • 期刊论文

  • 发表日期：

    2025年7月17日

查看全文
  • Sarika D. Patil,
  • Sarala P. Adhau,
  • Xma R. Pote,
  • Vaibhaw R. Doifode,
  • V. S. Rajguru,
  • Tejas R. Patil,
  • Ankita Jaiswal,
  • Nilesh Shelke,
  • Haytham F. Isleem

摘要

The artificial intelligence-based frameworks available to date have more or less failed to provide mechanical fidelity, resilience in real-world operating conditions, and multi-objective optimization capability. A holistic AI-based multi-criteria optimization framework developed in this work consists of Quantum Informed Graph Transformer Networks (QGTN), Self-Supervised Contrastive Multi-Objective Learning (SCMOL), Transformer-Based Evolutionary Surrogate Models (TESM), Physics Informed Generative Adversarial Networks (PI-GANs), and an AI-Driven Cyber-Resilient Digital Twin (AICR-DT). The proposed system augments the energy density of the baseline materials by 49.4%, while also improving cycle lifespan by 86.2%. The prediction accuracy of stability has improved to 94.7%, with a decrease of 97.6% in the generation of unstable structures. Moreover, the screening time by high throughput has been decreased by 85.9%, while real-time adaptive optimization has multiplied the supercapacitor life span by 3.1×. All these results reaffirm the robustness, scalability, and practicality of the proposed framework in accelerating the discovery and deployment of next-generation nanomaterials for high-performance electric vehicle (EV) supercapacitors.