Vibration and noise problems due to fluid flow occur in many industrial plants. This obstructs smooth plant operation. These flow-related phenomena are known as flow-induced vibrations (FIV). FIV is an important source of fatigue damage of mechanical equipments.

FIV can be divided into two kinds: with resonator and without resonator. One of the most famous FIV phenomena with resonator is cavity tone, which is a kind of acoustic resonance. Flow-induced acoustic resonance of cavities and side-branch pipes, are of high interest in engineering applications, such as the piping systems of power station, chemical plants, natural gas compressor stations, the fuel vents on aircrafts or vehicle conjunction of a train. One of the typical FIV phenomena without resonator is fluid flow inside a feedback fluidic oscillator. The feedback fluidic oscillator is a device which is characterized by the absence of moving parts and can be used to enact large-scale changes in a flowfield with a relatively small control input for aerodynamic flow control applications such as cavity resonance tone suppression, fluidic actuators, jet thrust vectoring and enhancement of jet mixing. Feedback fluidic oscillator has used internal flow separation and reattachment (Coanda effect) to generate the oscillations.

The present study focused on the frequency properties of FIV, i.e. clarifying lock-in condition and flow fluctuation propagation of FIV with resonator and frequency characteristics and flow oscillation pattern of FIV without resonator. In order to find out the shear layer or jet oscillation properties, particle image velocimetry (PIV) technique was utilized to extract the two-dimensional flow fields.

As a typical type of FIV with resonator, flow-induced acoustic resonances in a piping system containing closed coaxial side-branches with and without forced pressure fluctuation were investigated experimentally.

In order to excite a self-induced acoustic resonance in the piping system with closed coaxial side-branches (d/D = 1/4), a block was put at the inlet of the main pipe. For this piping system configuration, Resonance characteristics of the piping system were examined by a microphone firstly, as shown in Fig. 1 (a). The results revealed that the fluid flow fluctuations were strongly locked in corresponding to the natural frequencies of the side-branches for the flow under relative high Reynolds number condition (Re > 1.3 x 105). However, for the relative low Reynolds number flow conditions (6.6 x 104 < Re < 1.3 x 105), the resonance was seen to be dominated by the shear layer oscillation properties, i.e. the resonance occurred at a dominant frequency based on the shear layer oscillation frequency (frequency according to the hydrodynamic mode) rather than the natural frequency of the closed coaxial side-branches.

In order to induce much stronger acoustic resonance and make the resonance easy lock-in to the natural frequency of the coaxial resonator, a resonator with a width ratio of d/D = 4 in which acoustic radiation into the main pipe became smaller was designed and tested. Resonance characteristics of the piping system were examined by a microphone, as shown in Fig. 1 (b). In this kind of configuration, strong frequency lock-in was excited under the first acoustic resonance mode while Strouhal number St > 0.4. To increase the amplitude of the shear layer fluctuation and control the resonator operate under resonant and off-resonant conditions, a louder speaker was placed at the upper end side of the coaxial side-branches. Phase averaged velocity fields at eight successive 45°-wide interval phase of a typical acoustic cycle, were obtained two-dimensionally in the junction of coaxial side-branches, while the acoustic resonance was induced at the first and second hydrodynamic modes. Patterns of shear layer correspond to two hydrodynamic modes were obtained from the phase averaged velocity fields. The PIV can acquire time series velocity fluctuations, then, two-dimensional phase delay maps under resonance and off-resonance conditions in the junction of coaxial side-branches were obtained. Fig. 2 (a) (b) shows the contour map of phase difference under resonant condition, and Fig. 2 (c) shows the phase delay map under off-resonant condition. The phase delay maps under resonant condition were symmetric about the main pipe because the lower side-branch resonated with the speaker, pressure fluctuation passed down to the lower side-branch and made the two side-branches were well-tuned. For the phase delay maps under first and second hydrodynamic modes, the phase difference around the antinodes area of the shear layer became large dramatically. However, the contour lines in the junction under off-resonant condition exist only in the upper side-branch, where the loudspeaker was installed at the end, because the shear layer fluctuation disturbed the transformation of pressure fluctuation to the lower side-branch. Experimental results show that the proposed phase delay map method costs less experiment and computation time and achieves a better repetition than the phase locking technique. In addition, the phase delay map method can obtain phase difference under the different frequency components. This is extremely important when two different acoustic modes were induced in one experimental condition.

Motivating from understanding the mechanism of flow-induced vibration without resonator, characterization of periodic flow structure in a feedback fluidic oscillator under low Reynolds number water flow were investigated experimentally.

In the present study, a feedback fluidic oscillator has been manufactured and tested under low Reynolds number water flow, to clarify the flow patterns inside the oscillating chamber and the feedback channels, to figure out the oscillatory characteristics and dynamic structure of periodic jet fluctuation and the feedback flow, via PIV technique. The flow oscillations were triggered by the Coanda effect, the frequency highly increases with the Reynolds number and a non constant Strouhal number while Reynolds number ranged from 200 to 630 were obtained based on spectral analysis of the velocity fluctuations inside the feedback channels, as shown in Fig. 3 (a). The flow rate fluctuations of the feedback flows inside the feedback channels were quantitatively determined. Meanwhile, corresponding patterns of the jet fluctuation during one cycle were extracted from the two-dimensional velocity distributions. The time of jet oscillate from the center of the control volume to either of the attachment walls and that of jet oscillate from either of the attachment walls to the center of the control volume were separated and discussed by different interpretation models. The experimental results showed that the time of jet oscillate from the center of the control volume to either of the attachment walls was related to the feedback flow through the feedback channels and the time of jet oscillate from either of the attachment walls to the center of the control volume was related to the propagation time of jet traveling along the oscillating chamber, as shown in Fig. 3 (b), and this time was found approximately equal to 5 times propagation time of jet traveling along the oscillating chamber under low Reynolds water flow.

Fig. 1 Resonance characteristics of closed coaxial side-branches system. Dashed lines correspond to hydrodynamic mode, m = acoustic modes number. (a) d/D = 1/4, (b) d/D = 4.

(a) Under 1st hydrodynamic mode (b) Under 2nd hydrodynamic mode (c) Under off-resonance condition

Fig. 2 Phase delay maps under resonant and off-resonant conditions.

Fig. 3 Frequency characteristics of the feedback fluidic oscillator (a) and characteristics time of the jet

oscillation. ▲ shows the half oscillation period. ● indicates the time of jet oscillation from the center of the control volume to one of the attachment walls. ▼ represents five times jet oscillation time. ◆ signifies the time of jet oscillation from one of the attachment walls to the center of the control volume.

本論文は,大型プラントや新幹線・自動車といった高速移動物体に関するトラブル要因となるなど工学的重要性が高い流体関連振動現象に着目し,最先端の実験的手法によってこれらの流体現象に関する詳細データを取得し,振動発生条件や振動数決定機構を解明することを目的としてデータ分析・検討を行った成果について述べられている.

本論文では流体の自励振動現象を,流れ場に固有振動数を持つ物体が存在し,その固有振動数への引き込み現象を伴う可能性のある振動現象と,特定の固有振動数が存在しない系で生じる振動現象とに大別して研究を進めている.本論文では,キャビティトーン現象を前者の代表的な振動現象として,小型フルイディックスを後者の代表的な振動現象として取り上げ,最先端の実験的手法によって精査している.キャビティトーン現象は,人工物の騒音振動やエネルギーシステムの高出力化に随伴して生じる音響共鳴など,多くの人間環境システムで問題となっている.古くから工学的利用が進められている流体振動子は,最もシンプルな流体自励振動であり,小型省エネ機器への組み込みが試みられているが,小型システム内では発振しにくいなどの問題がある.

第一章では固有振動数が系に存在する場合としない場合の振動現象を広く調査し,自励振動に必要なフィードバックシステムのこれまでの定説についてまとめ,本研究の目的を述べている.第二章では,キャビティトーン現象,フルイディックスについてそれぞれ既往の研究を詳細にまとめ,これら既往の研究に欠けている点を指摘している.

第三章ではキャビティトーン現象発生時の流れの挙動を高速度PIV実験によって明らかにした.実験では十字共鳴管を持つキャビティについて,その振動発生条件と密接に関連するロックイン現象に着目し,ストローハル数0.4以下で流れが振動する比較的低レイノルズ数条件(約10万以下)においては,共鳴管の固有振動数から50%離れた振動数条件であっても,ストローハル数一定となる振動が生じること,10万以上のレイノルズ数条件では強いロックインが生じて気柱の固有振動数で振動が発生することを明らかにした.また,振動発生時の振動位相の遅れを空間的に測定する手法を提案するとともにこれを明らかにした.ロックイン現象は構造物と流れの相互作用による自励振動現象を含め,固有振動系の存在する多くの流れ場で生じるが,そのロックイン現象について,ストローハル数とレイノルズ数によって,発生の有無を明らかにしている.また,振動の空間的伝播を求める手法は汎用的であり,有用な工学的手法である.

第四章では,フィードバックループ付きの小型フルイディクスの振動数決定モデルについて検討した結果をまとめている.小型フルイディクスは工学的な利用が進められている通常のフルイディクスと異なり,ストローハル数が一定ではない,振動数が流速に比例しない特異な挙動を示す.このメカニズムや振動数決定メカニズムは不明であり,小型のエネルギー環境機器への導入が進んでいない.本論文では振動発生時の流れ場全体の速度分布を捉え,局所的な速度差が100倍以上もある流れの分析を,高速度PIVを用いて行った.また,汎用数値シミュレーションコードを用いた数値的な検討も行っている.実験の結果に基づき,フィードバックループ内の流れが,振動を噴流が直進方向から容器壁まで移動するのに要する時間を支配しており,液体流れでありながら圧力フィードバック系とは異なる点で既存の大型システムと異なることを明らかにした.また,容器壁から直進方向まで噴流が立ち直るのに要する時間は,流れの擾乱の伝播速度,ストローハル数で決まっていることを明らかにした.更に,小型フルイディクスでは,これら二つのプロセスとコアンダ効果との相互作用によって噴流が容器壁に留まる,二つのプロセス間の遷移が時間遅れを伴うことを明らかにし,新しい振動数決定モデルを提案した.これはフルイディクスの小型環境エネルギー機器への導入に重要な一般的な知見である.

審査では,以上の内容について,特に小型フルイディクスの振動プロセスに関する物理的な意味に着目した議論が行われた.その結果,本研究において新しく提案した実験的な手法や提案した振動モデルの学術的意義は,博士(環境学)の学位を受けるにふさわしいとの審査結果に至った.

したがって,博士(環境学)の学位を授与できると認める.