1.本募集の概要：

　Office of Naval Research-Global (ONRG)の課題
（課題番号：N62909-24-1-2117、契約期間：2024年9月から2027年9月)
　を遂行するためにポスドクを募集する。

背景と課題の簡単な説明は以下の通り。

The investigation of radar cross section (RCS) applications is a long-standing and open defense technology problem. Actually, any RCS prediction for a realistic object of complex structure and material composition, like an aircraft, depends inevitably on numerical computations. Recently, due to the use of advanced configurations and media in modern aircrafts, the RCS analysis must handle even more complicated wave phenomena. To address such an issue, highly precise RCS calculation tools are required in order to simulate low-observable and electrically-large sized aircraft designs. The NS-FDTD method, proposed by J. Cole in 1995 [1], is a noteworthy option that reduces the dispersion error significantly, often observed in other techniques, like the FDTD one. In essence, the NS-FDTD algorithm offers extremely high-precision CW analysis by using elaborately devised models and highly isotropic finite-difference operators, based on the Laplacian concept. Its calculation accuracy is 4 orders of magnitude higher than that of the FDTD approach for the same cell size. Moreover, the NS-FDTD method can easily and reliably handle arbitrarily-shaped objects with several media, unlike existing realizations.

The FDTD technique is often used for the RCS prediction of aircrafts, despite its significant dispersion errors. Typically, an aircraft has an electrical size of around 500λ-1800λ, where λ is the radar wavelength. Unfortunately, along the axial direction, the corresponding FDTD wavelength reaches the level of 7λ-25λ; values that are unacceptable in the design of modern air vessels. On the other hand, the corresponding error of the NS-FDTD scheme is zero. Despite this advantage, however, the NS-FDTD method is formulated according to a differential form, applied to discrete points on orthogonal grids. Hence, its modeling accuracy can greatly degrade in the case of real-world objects, whose curved shape or intricate interfaces are not aligned to the orthogonal grid of a computational domain.

To overcome the prior serious difficulties, the simplest way could be the use of very fine grids, nonetheless at the expense of unduly large and practically unaffordable computational overheads. Lately, an efficient alternative, which launches a contour-path (CP) model via the integral form of Maxwell’s equations, has been reported for the FDTD method [2]. Although promising, the specific CP model (in its original form) is still not the ideal solution for RCS problems, since it would, still, have large memory requirements, especially for the nonmetal parts of an aircraft.

Based on these aspects, the goal of this project is to develop a versatile and robust subgridding algorithm for electrically-large objects, like the radome of an aircraft fuselage. In fact, we currently investigate a simpler subgridding scheme, based on 3-D CNS-FDTD models [3]. The key steps of the proposed approach are discussed below.

We have conducted extensive and leading research on the development of the NS-FDTD method since 1999. Evidently, the RCS analysis is one of the most important areas in electromagnetic wave propagation. Since our initial involvement, we have proposed several new techniques to enhance and facilitate practical NS-FDTD calculations for electrically-large objects with complicated structures and dielectrics; such as the complex-type NS-FDTD scheme as well as the surface impedance boundary condition (SIBC) technique and path integral (PI) concept for the NS-FDTD formulation.

In this project, we intend to introduce a simple 3-D subgridding technique to treat objects with an electrical size of 60-70. Since we do not have an RCS experimental facility, we plan to validate our technique by comparing our results with those published in the relevant literature. Basically, we plan to conduct our research according to the following stages:

1)      Basic subgridding scheme: The goal in this stage is to construct a basic subgridding algorithm to analyze electrically-large objects, like an aircraft radome. As of February, 2024, we are investigating a modeling technique, based on the 3-D CNS-FDTD subgridding system [3]. The main issue is the numerical instability and divergence in the case of long iterations along with wave distortion at the connection of different cells. Firstly, we pursue a stable connection algorithm and next, we try to reduce the wave distortion at the interface of different cells. Note that, even at this stage, the calculation speed is a serious concern.

2)      Acceleration via GPU: As all problems should be solved at a rational time, the acceleration of our method is another important task, particularly for large-scale calculations. Generally, modern fighters are electrically-large structures and the presence of dielectric materials requires extremely detailed and computationally demanding (CPU and RAM) simulations, due to the stability condition of the NS-FDTD method. To solve this issue, we will explore a speed up by applying a GPU based on our previous research [4].

3)      Stability and performance verification: The overall accumulation error of the featured technique is of utmost importance and therefore it constitutes our highest priority during this stage, since the numerical analysis of large-scale problems leads to prolonged simulations. For this reason, we meticulously investigate the stability and performance of the proposed subgridding algorithm of 3-D realistic aircraft models, taking into account the scattering waves at radar frequencies.

References

[1]     J. B. Cole, “A high accuracy FDTD algorithm to solve microwave propagation and scattering problems on a coarse grid,” IEEE Trans. Microw. Theory Tech., vol.43, no. 9, pp. 2053-2058, 1995.

[2]     T. Ohtani, Y. Kanai, and N. V. Kantartzis, “A nonstandard path integral model for curved surface analysis,” Energies, vol. 15, art. no. 4322, 2022.

[3]     Y. Kanai, T. Hoshino, T. Ohtani, and N. V. Kantartzis, “GPU acceleration of the nonstandard FDTD method”, 2023 Applied Computational Electromagnetics Society (ACES) Symposium, Monterey, CA, USA, Mar. 26–30, 2023.

[4]     T. Ohtani, Y. Kanai, and N. V. Kantartzis, “An enhanced stability subgrid scheme for the nonstandard-FDTD technique,” in Proc. 2025 Applied Computational Electromagnetics Society (ACES) Symposium, Orlando, FL, USA, May 2025.

2.募集する職種及び募集人員：

　研究員（ポスドク）　1名

　

3.勤務地：

　新潟工科大学（新潟県柏崎市藤橋1719番地）もしくは自宅

4.所属：

　新潟工科大学　金井フェロー研究室

5.職務内容：

１）NS-FDTDに基くサブグリッド法計算式のプログラム化（CPUおよびGPU)
２）計算精度の評価、計算の高速化
３）英文による報告書および発表資料の作成

6.資格･条件：

　以下のいずれかの方
　・博士の学位を有する方
　・着任時までに博士の学位を取得見込みの方
　・電磁界の数値解析、特にFDTD法、の研究に従事してきた方

7.採用条件

　令和7年7月1日以降、なるべく早い時期（応相談）

8.雇用期間：

　採用日から令和8年3月31日まで
　※以下の場合、任期途中であっても解雇する場合がある。
　１）当該の研究費の残高が不足した場合
　２）当該の研究、事業が終了した場合
　３）受入先研究室のフェローの任期が終了した場合
　４）能力不足、勤務態度不良があった場合

9.勤務日：

　週5日（月曜～金曜）

10.勤務時間：

　専門業務型裁量労働制（1日8時間勤務したとみなす）

11.休日、休暇：

　土曜、日曜、祝日、夏季特別休業、年末年始休暇他

12.給　与：

　・本　給：月額給与：30万円（学位取得後1年未満）、35万円（同1年以上3年未満）40万円（同3年以上）
　・諸手当：通勤手当（本学規程により支給）
　・支給日：毎月末締め、原則当月25日払い

13.社会保険：

　・日本私立学校振興・共済事業団（健康保険、年金）
　・雇用保険
　・労災保険

14.提出書類：

　18．記載のメールアドレスに以下の書類を添付し、件名を「研究員採用選考応募」として送付してください。

　1. 履歴書：様式任意（写真貼付、e-mailアドレスを含む連絡先と着任可能時期を明記）
　2. 研究業績一覧：様式任意（論文／発表種別、査読有無で分類して記載）
　3. 主要な学術論文のコピー（3編以内）
　4. これまでの研究概要と研究の抱負：様式任意（A4用紙1枚程度）
　5. 推薦書または照会可能者2名の氏名と連絡先

　＊応募書類は返却しません。応募書類に含まれる個人情報は本人事選考にのみ使用し、他の目的には一切使用しません。

15.選考方法：

　書類選考の後、必要に応じて面接日をご連絡します。面接は原則オンラインで行う予定ですが、対面で面接試験等を行う場合、来所に係る旅費等は応募者の負担とします。

16.採否通知：

　採否については、速やかに本人あてに通知します。

17.応募締切：

　令和7年12月31日（水）必着

延長の可能性あり。ただし、随時選考を行い、適任者が決定次第締切ります。

18.書類送付先・問い合わせ先：

　[書類送付先・研究内容等問い合わせ先]
　〒945-1195　新潟県柏崎市藤橋1719番地　新潟工科大学　
　新潟工科大学フェロー　金井　靖　宛　
　TEL：0257-22-8111
　Email：kanai@iee.niit.ac.jp