2022 Fiscal Year Final Research Report
Characteristic Elucidation and Shape Optimization of Multiconductor Transimission Lines by Using Isogeometric Analysis and Machine Learning
Project/Area Number |
19K14961
|
Research Category |
Grant-in-Aid for Early-Career Scientists
|
Allocation Type | Multi-year Fund |
Review Section |
Basic Section 21010:Power engineering-related
|
Research Institution | Shizuoka University |
Principal Investigator |
|
Project Period (FY) |
2019-04-01 – 2023-03-31
|
Keywords | 多導体ケーブル / アイソジオメトリック解析 / 特性解明 / 形状最適化 / 重回帰分析 / 最適化アルゴリズム / 機械学習 / 深層学習 |
Outline of Final Research Achievements |
In this research, the transmission and radiation characteristics of multiconductor cables consisting of multiple conductors such as signal lines, power lines, and ground lines were clarified, and a shape optimization technique was developed to design cables with good transmission characteristics and low unwanted radiation. Specifically, the transmission and radiation characteristics of multiconductor cables were first calculated efficiently and accurately by "electromagnetic field simulation based on isogeometric analysis. Then, by combining the calculated characteristics with "machine learning" to generate clustering and regression models, we verified and clarified the effects of cable shape on transmission/radiation characteristics, and created a design technique to derive the optimal cable shape.
|
Free Research Field |
数値シミュレーション
|
Academic Significance and Societal Importance of the Research Achievements |
多導体ケーブルをに関するこれまでの手法では,多導体ケーブルの滑らかな曲線形状を精度良く再現できているとは言えない.そのため,アイソジオメトリック解析におけるNURBS曲線・曲面を多導体ケーブルのモデル化に用いるような研究は本研究を除いて存在しない.また,多導体ケーブルの根本的な特性の解明を試みたものはほとんどなく,理論的な解明はまだなされていない.加えて,特にEMCの分野において機械学習を応用した研究はほとんどないため,機械学習に基づく検証方法や最適化手法を創出したことによって,学術的にも社会的にも国内外に大きなインパクトを与えることが期待できる.
|