Scholarlink

<p>Optimizing gear geometry is a&#xa0;challenging and resource-intensive task, primarily due to the interdependent and distinct nature of various design parameters. Transmission systems require custom designs tailored to specific applications with the primary goals of maximizing durability, minimizing noise (NVH), reducing cost, and increasing efficiency of the gears. These often-conflicting goals create a&#xa0;complex optimization problem that has traditionally required considerable time and expertise.</p><p>To address these challenges, an automated, AI-integrated tool has been developed that significantly streamlines and accelerates the geometry optimization process of spur gears through iterative, data-driven refinement. The solution includes an adaptive spur gear geometry generator, which enables flexible configurations based on established geometry requirements. A&#xa0;custom AI solver, trained on application-specific data, models the durability, NVH, cost, and efficiency characteristics of each gear configuration. To ensure high modeling accuracy and computational speed, a&#xa0;framework for the automated search of optimal hyperparameters for the AI solver is presented. A&#xa0;performance scoring function evaluates and ranks each geometry based on predefined metrics. These components are coordinated by optimization logic that iteratively adjusts design parameters, generating and refining successive configurations. This collaborative framework enables rapid and precise exploration of the gear design space, leading to optimal solutions with exceptional efficiency.</p><p>Evaluation of the proposed automated model development framework demonstrates its effectiveness, achieving high modeling accuracy and significantly reducing computation time. The application of the design tool to an e‑bike transmission use case validates its capability to generate optimized gear geometries while adhering to design constraints, meeting durability and NVH requirements, and optimizing costs. These results highlight the potential of the methodology for holistic transmission design, enabling systematic optimization of multi-stage gear systems.</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

Application of artificial intelligence tools for the automated optimization of gear geometries

  • 期刊论文

  • 发表日期：

    2025年8月22日

查看全文
  • Tobias Jonathan Boschmann,
  • Philipp Kauffmann

摘要

Optimizing gear geometry is a challenging and resource-intensive task, primarily due to the interdependent and distinct nature of various design parameters. Transmission systems require custom designs tailored to specific applications with the primary goals of maximizing durability, minimizing noise (NVH), reducing cost, and increasing efficiency of the gears. These often-conflicting goals create a complex optimization problem that has traditionally required considerable time and expertise.

To address these challenges, an automated, AI-integrated tool has been developed that significantly streamlines and accelerates the geometry optimization process of spur gears through iterative, data-driven refinement. The solution includes an adaptive spur gear geometry generator, which enables flexible configurations based on established geometry requirements. A custom AI solver, trained on application-specific data, models the durability, NVH, cost, and efficiency characteristics of each gear configuration. To ensure high modeling accuracy and computational speed, a framework for the automated search of optimal hyperparameters for the AI solver is presented. A performance scoring function evaluates and ranks each geometry based on predefined metrics. These components are coordinated by optimization logic that iteratively adjusts design parameters, generating and refining successive configurations. This collaborative framework enables rapid and precise exploration of the gear design space, leading to optimal solutions with exceptional efficiency.

Evaluation of the proposed automated model development framework demonstrates its effectiveness, achieving high modeling accuracy and significantly reducing computation time. The application of the design tool to an e‑bike transmission use case validates its capability to generate optimized gear geometries while adhering to design constraints, meeting durability and NVH requirements, and optimizing costs. These results highlight the potential of the methodology for holistic transmission design, enabling systematic optimization of multi-stage gear systems.