betaFeedback

WorksPhysics and Astronomy

Pressure fluctuations of liquids under short-time acceleration

Abstract

This study experimentally investigates the pressure fluctuations of liquids in a column under short-time acceleration and demonstrates that the Strouhal number StSt [=L/(cΔt)=L/(cΔt), where LL, cc, and ΔtΔt are the liquid column length, speed of sound, and acceleration duration, respectively] provides a measure of the pressure fluctuations both for limiting cases (i.e. St≪1St\ll1 or St=∞St = \infty) and for intermediate StSt values. Incompressible fluid theory and water hammer theory respectively imply that the magnitude of the averaged pressure fluctuation P‾\overline{P} becomes negligible for St≪1St\ll1 (i.e., in the condition where the duration of acceleration ΔtΔt is large enough compared to the acoustic timescale) and tends to ρcu0ρcu_0 (where u0u_0 is the change in the liquid velocity) for St≥O(1)St\geq O(1) (i.e., in the condition where ΔtΔt is small enough). For intermediate StSt values, there is no consensus on the value of P‾\overline{P}. In our experiments, LL, cc, and ΔtΔt are varied so that 0.02≤St≤2.20.02 \leq St \leq 2.2. The results suggest that the incompressible fluid theory holds only up to St∼0.2St\sim0.2 and that StSt governs the pressure fluctuations under different experimental conditions for higher StSt values. The data relating to a hydrogel also tend to collapse to a unified trend. The inception of cavitation in the liquid starts at St∼0.2St\sim 0.2 for various ΔtΔt, indicating that the liquid pressure becomes negative. To understand this mechanism, we employ a one-dimensional wave propagation model with a pressure wavefront of finite thickness that scales with ΔtΔt. The model provides a reasonable description of the experimental results as a function of StSt. The slight discrepancy between the model and experimental results reveals additional contributing factors such as the container motion and the profile of the pressure wavefront.

Introduction

Abrupt liquid motion causes pressure fluctuations inside the liquid column, resulting in significant fluid motion. For example, a large deformation of the free surface (i.e. jets) can occur in Pokrovski’s experiment, where a liquid-filled container falls under gravity and hits the floor [2]. The pressure impulse approach under the incompressible fluid assumption explains the jet formation. Nevertheless, Kiyama et al. [25] found that sufficiently large acceleration causes cavitation accompanied by jet formation with a vibrating interface. This implies that the acoustic pressure waves propagate in the liquid column [3], meaning that the jet liquid behaves as a weakly compressible fluid.

To describe the liquid compressibility, scaling analysis [5 §6.3, p. 168-] provides three important terms related as

∣1ρc2∂p∂t−12c2Dq2Dt+aUc2∣≪UL,(1)\left|\frac{1}{\rho c^{2}}\frac{\partial p}{\partial t}-\frac{1}{2c^{2}}\frac{Dq^{2}}{Dt}+\frac{aU}{c^{2}}\right|\ll\frac{U}{L}, \tag*{(1)}

where ρ\rho is the liquid density, cc is the speed of sound, pp is the pressure, qq is a quantity having the same dimensions as velocity, aa is the typical acceleration, UU is the typical velocity of liquid (e.g., the impact velocity in an example of Pokrovski’s experiment), and LL is the typical length scale of the liquid system. Note that the flow is assumed to be isentropic. When each term on the left-hand side has a much smaller magnitude than the spatial derivatives of the components of liquid velocity U/LU/L, the liquid behaves as if it were incompressible. This consideration gives three dimensionless incompressibility conditions relating to the Mach number MaMa, Strouhal number StSt, and Froude number FrFr, defined as

LcΔt=St≪1,(2)\frac{L}{c\Delta t}=St\ll1, \tag*{(2)}
Uc=Ma≪1,(3)\frac{U}{c}=Ma\ll1, \tag*{(3)}
aLc2=Fr−1≪1,(4)\frac{aL}{c^{2}}=Fr^{-1}\ll1, \tag*{(4)}

where Δt\Delta t is the duration required for the development of the pressure field. The Strouhal number StSt is the ratio of the duration of acceleration Δt\Delta t to the acoustic timescale L/cL/c; a detailed explanation will be provided in § 3.1. Note that the Froude number FrFr is generally expressed as the ratio of the inertial force to gravitational force. The physical meaning of FrFr in the above equation is the competition between the pressure change due to the body force (acceleration including gravity) ∼ρaL\sim\rho aL and the ambient pressure ∼ρc2\sim\rho c^{2}. This dimensionless number aL/c2aL/c^{2} is equivalent to Fr−1Fr^{-1}, and is also the product of the other two dimensionless numbers (i.e. Fr−1∼StMaFr^{-1}\sim StMa) when the typical acceleration can be scaled as a∼U/Δta\sim U/\Delta t.

In a previous example of Pokrovski’s experiment with fluid compressibility effects [25], the typical values were found to be U∼O(100)U\sim O(10^{0}) m/s, c∼O(103)c\sim O(10^{3}) m/s, L∼O(10−1)L\sim O(10^{-1}) m, Δt∼O(10−4)\Delta t\sim O(10^{-4}) s, and a∼O(104)a\sim O(10^{4}) m/s2^{2}. Under such conditions, Ma∼O(10−3)≪1Ma\sim O(10^{-3})\ll1, St∼O(100)St\sim O(10^{0}), and Fr−1∼O(10−3)≪1Fr^{-1}\sim O(10^{-3})\ll1. This suggests that the compressibility effect in Pokrovski’s experiment might be scaled with the Strouhal number StSt, rather than with MaMa and FrFr. However, Kiyama et al. [25] did not consider the influence of the acceleration duration (i.e. StSt).

We tested the above hypothesis in a preliminary experiment. Figure 1 shows the time series of the acceleration aa measured at the top of a container filled with silicone oil (10 cSt) as it impacts the floor. The red line indicates the data taken from the tube impacting a metal floor, where the duration of acceleration Δt∼0.11\Delta t\sim0.11 ms (indicated by the red shading). The blue line presents the data for a resin floor (Δt∼0.27\Delta t\sim0.27 ms, blue shading). The dimensionless parameters are St∼3.1×10−1St\sim3.1\times10^{-1}, Ma∼8.2×10−4≪1Ma\sim8.2\times10^{-4}\ll1, and Fr−1∼3.6×10−5≪1Fr^{-1}\sim3.6\times10^{-5}\ll1 (red) and St∼1.2×10−1St\sim1.2\times10^{-1}, Ma∼2.1×10−3≪1Ma\sim2.1\times10^{-3}\ll1, and Fr−1∼3.6×10−5≪1Fr^{-1}\sim3.6\times10^{-5}\ll1 (blue). Although both cases exhibit similar peaks and mean accelerations at impact, the magnitude of later acceleration fluctuations is quite different. The case with the larger Strouhal number StSt (i.e. a smaller Δt\Delta t) exhibits more significant acceleration fluctuations for t>Δtt>\Delta t. This confirms that the short-term acceleration induces pressure waves in the liquid column, even when Ma≪1Ma\ll1 and Fr−1≪1Fr^{-1}\ll1. We conjectured that the influence of Δt\Delta t (and thus StSt) became visible there, as the other two dimensionless parameters (MaMa, Fr−1Fr^{-1}) remain much smaller than unity, implying that they are less important. The effects of the acceleration duration Δt\Delta t are also visible in the fluid motion inside test tubes dropped from the same height (i.e. the same velocity UU, see Figure 2). The resin floor induces a smooth jet (Figure 2a), while the metal floor triggers surface vibrations (Figure 2b) and cavitation (Figure 2c); see supplementary movies.

Image not extracted

Figure 1. Measured acceleration of a liquid-filled glass container after collision with the floor. Red and blue curves show the acceleration with different floor materials and drop heights. The acceleration during the impact is marked by the shaded area. Although both cases have similar peaks and mean accelerations at impact, the magnitude of the subsequent acceleration fluctuations is quite different.

Image not extracted

Figure 2. Photographs taken at 100,000 fps using a high-speed camera (Photron SA-X) with a back-light method. The typical velocity is mostly constant (≈2.0\approx2.0 m/s) and the height of the jet does not change significantly in each case. The rightmost panels show a magnified view of the free surface at t=4.6t = 4.6 ms. (a) Test tube impacting a resin floor, where the free surface remains smooth. (b) Test tube impacting a metal floor, where the free surface vibrates periodically and exhibits a rough texture (marked by a white arrow). (c) Test tube impacting a metal floor, where cavitation occurs inside the liquid. The free surface exhibits a similar response to that in (a). See supplementary movies.

To the best of our knowledge, the role of Δt\Delta t in the development of the pressure field in Pokrovski’s experiment has not been systematically investigated. Most existing research took a pressure impulse approach assuming an incompressible flow and a constant Δt\Delta t which is greater than the acoustic timescale L/cL/c, i.e. St≪1St \ll1 ([2, 34]). In this approach, the pressure field of the liquid is fully developed immediately after the impact, and thus the liquid pressure does not change as a function of time. Another approach takes the water hammer theory into account ([25]), and thus assumes an instant increment in pressure, i.e. Δt=0\Delta t = 0 ([15, 6]), where St=∞St = \infty. This assumption predicts periodic pressure fluctuations PP with a magnitude of ρcU\rho c U, but does not capture the role of Δt\Delta t in the development of the pressure field. Existing research has considered one or other of these approaches, but the intermediate region between St≪1St \ll1 and St=∞St = \infty has rarely been studied.

This paper focuses on the role of StSt in the pressure fluctuations, especially in the intermediate StSt regime. We first define the physical meaning of StSt in § 3.1 as a function of Δt\Delta t, and examine the experimental data with various liquid depths LL, velocities UU, acceleration durations Δt\Delta t, and liquid types. The experimental data collapse onto a single curve, suggesting that StSt is suitable for describing the pressure fluctuations (§ 3.2). The conditions for the onset of cavitation are derived in § 3.2.2 and the pressure fluctuations in a hydrogel are examined in § 3.2.3. In both cases, the present results are shown to be in line with existing pressure results for fluids. We also develop a simple model which takes the effect of the finite thickness of the pressure wavefront into account. This model is used to describe the liquid pressure fluctuations at t>Δtt > \Delta t based on the one-dimensional wave equation (§ 3.3). We then compare the output from the proposed model with a wide range of experimental data. The influence of the motion of the surrounding container and the profile of the pressure wavefront is also discussed.

Experiments

Figure 3 shows the experimental setup. The container, which is partially filled with a liquid/hydrogel, falls freely and eventually collides with the floor. This accelerates the liquid in the vertical direction opposite to the direction of gravity gg. The acceleration of the container is measured with an accelerometer [2350, Showa Sokki Co., sensibility 0.3 pC/(m/s2^2)] attached to the top of the container. The accelerometer outputs a charge which is converted to a voltage by a charge amplifier (5015A, 5011B, Kistler Co.), and the voltage is recorded by an oscilloscope (Iwatsu Co., ViewGo II, DS 5554-A).

Image not extracted

Figure 3. Schematic illustration of the experimental setup. The accelerometer attached to the top of the container records the acceleration aa acting on the system. A change in the floor material allows us to change the duration of acceleration Δt\Delta t. The depth of the liquid in the container LL is also varied in our experiments.

The control parameters are the depth of the liquid column LL, the change in the liquid velocity induced by the impact u0u_0, the duration of acceleration Δt\Delta t, and the type of liquid/hydrogel. The depth of the liquid column LL ranges from 30–285 mm. We use two borosilicate glass containers of different sizes and masses (see Table 1). The characteristic velocity UU in this experiment corresponds to the impact velocity u0u_0, which is varied from 0.5–3 m/s by adjusting the drop height of the container. The duration of acceleration Δt\Delta t ranges from 0.1–2.2 ms. To control Δt\Delta t, different floor materials are used: steel, aluminium, epoxy resin, ABS resin, and rubber. We use different fluids (silicone oils with 1 and 10 cSt obtained from Shin-Etsu Chemical Co., pure water, and ethanol) and a weak gelatin gel (5 wt%) that is expected to flow when it experiences rapid deformation ([24]) and exhibits similar cavitation to water ([35]). The physical properties of these media are summarized in Table 2.

TubeLength [mm]Weight [g]Inner diameter [mm]Wall thickness [mm]
Short13019141
Long30066122

Table 1. Specifications of the containers.

MediaSpeed of sound cc [m/s]Density ρ0\rho_0 [kg/m3^3]
Silicone oil (10 cSt)966.5935
Silicone oil (1 cSt)901.3818
Water1483998
Ethanol1168789
Gelatin13911052

Table 2. Physical properties of media in a tube. Speed of sound in gelatin is calculated from the acceleration frequency and liquid depth.

A measured acceleration profile is shown in Figure 4(a). A positive value of the acceleration indicates upward vertical acceleration. The grey curve shows the raw acceleration data, whereas the red curve shows the low-pass-filtered data. For filtering, the cut-off frequencies for the 130 and 300 mm tube lengths are 11,000 and 5,000 Hz, respectively; these values are estimated based on the tube length and the speed of sound in glass (c∼5.4×103c \sim5.4 \times10^{3} m/s). The raw data suddenly change at t=0t = 0 when the container impacts the floor, and the acceleration exceeds 1,000 gg immediately after the impact. We define Δt\Delta t as the duration from the collision until the filtered acceleration falls back to 0 m/s2^{2}. Periodic fluctuations in acceleration are then observed (see Figure 4 for t>t0t > t_0). At t>t0t > t_0, no fluctuations are visible for an empty container (see inset), indicating that there must be some liquid/hydrogel in the container to observe fluctuations.

Image not extracted

Figure 4. (a) Acceleration of the container partially filled with silicone oil (10 cSt). The grey and red lines represent raw data and data subjected to a low-pass filter. Δt\Delta t is the duration of acceleration. After t0t_0 (defined in §3.4), fluctuations in acceleration are visible. The inset shows the acceleration of an empty glass container. (b) Acceleration of each floor type is shown by different colours. Data are obtained from the impact of an empty test tube on each floor material. The Young’s modulus is specified in the legend, and * denotes the tensile strength.

The acceleration during Δt\Delta t for each floor type is shown in Figure 4(b). The container is dropped from the same height so that the integral of acceleration over Δt\Delta t is similar for all cases. However, a stiffer floor introduces a higher magnitude of acceleration aa within a shorter Δt\Delta t.

The measured acceleration is translated to pressure according to the momentum conservation law: the amplitude of momentum change in the total liquid, ΠL\Pi_L, is equal to that of the container, ΠT\Pi_T. The amplitude of the momentum change in the total liquid is expressed as

ΠL=ρA(∫0Lu dx)max⁡−(∫0Lu dx)min⁡2=ρALU‾,(5)\Pi_L=\rho A\frac{\left(\int_0^L u\,\mathrm{d}x\right)_{\max}-\left(\int_0^L u\,\mathrm{d}x\right)_{\min}}{2}=\rho A L\overline{U}, \tag*{(5)}

where U‾\overline{U} represents the amplitude of the time difference and the spatial average of the velocity of the liquid, and AA indicates the cross-sectional area of the liquid column. The amplitude of the change in momentum of the container is expressed as

ΠT=mV,(6)\Pi_T=mV, \tag*{(6)}

where mm and VV are the mass of the container (see Table 1) and the amplitude of the velocity fluctuations of the container, respectively. Coupling equations (5) and (6) gives the following expression:

U‾u0=mρALVu0.(7)\frac{\overline{U}}{u_0}=\frac{m}{\rho A L}\frac{V}{u_0}. \tag*{(7)}

We can experimentally determine the right-hand side of equation (7). We estimate the velocity of the container VV by integrating the acceleration data (see Figure 4a). We focus on the velocity after t=t0t=t_0 and detect the velocity fluctuations vmax⁡v_{\max} and vmin⁡v_{\min} (see Figure 5). Note that vmax⁡v_{\max} and vmin⁡v_{\min} are taken from the first period after t=t0t=t_0 to reduce the influence of drift noise. The magnitude of the fluctuations in VV is defined as the mean value of vmax⁡v_{\max} and vmin⁡v_{\min}. Assuming the relationship P‾=ρcU‾\overline{P}=\rho c\overline{U}, we obtain

Image not extracted

Figure 5. Temporal velocity of the container (i.e. test tube) from the beginning of impact, as calculated by integrating the acceleration data. The peak velocities vmax⁡v_{\max} and vmin⁡v_{\min} are marked.

U‾u0=P‾ρcu0.(8)\frac{\overline{U}}{u_0}=\frac{\overline{P}}{\rho c u_0}. \tag*{(8)}

Hereafter, the experimental data P‾/(ρcu0)\overline{P}/(\rho c u_0) are estimated as the representation of mV/(ρALu0)mV/(\rho A L u_0).

Results and discussion

Strouhal number St as an indicator of liquid compressibility

We first derive the Strouhal number StSt for the one-dimensional system. We assume an infinitesimal disturbance to the velocity, pressure, and density in the inviscid flow. We then obtain the continuity and Euler equations in terms of the density ρ\rho, velocity uu, time tt, length xx, and pressure pp as

∂ρ∂t+ρ0∂u∂x=0,(9)\frac{\partial\rho}{\partial t}+\rho_0\frac{\partial u}{\partial x}=0, \tag*{(9)}
∂u∂t+1ρ0∂p∂x=0.(10)\frac{\partial u}{\partial t}+\frac{1}{\rho_0}\frac{\partial p}{\partial x}=0. \tag*{(10)}

Here, ρ0\rho_0 is the characteristic liquid density in the far field. The speed of sound in the liquid, cc, is defined as c2=∂p/∂ρc^2=\partial p/\partial\rho. Coupling equations (9) and (10) yields the wave equation [28] in the form

∂2Φ∂t2−c2∂2Φ∂x2=0,(11)\frac{\partial^2\Phi}{\partial t^2}-c^2\frac{\partial^2\Phi}{\partial x^2}=0, \tag*{(11)}

where Φ\Phi is the velocity potential [13]. Introducing dimensionless quantities, the wave equation can be written as

St2∂2Φ∗∂t∗2−∂2Φ∗∂x∗2=0,(12)St^2\frac{\partial^2\Phi^*}{\partial t^{*2}}-\frac{\partial^2\Phi^*}{\partial x^{*2}}=0, \tag*{(12)}

where St=x/(ut)St=x/(ut) is the Strouhal number. The superscript * denotes dimensionless quantities. In this work, the length of the liquid column LL, speed of sound cc, and duration of acceleration Δt\Delta t are substituted for the characteristic length xx, speed uu, and time tt, respectively. The Strouhal number StSt for our experiments is thus defined as

St=LcΔt,(13)St=\frac{L}{c\Delta t}, \tag*{(13)}

which is the same form as equation (2). In this form, StSt can be considered as the ratio between the characteristic length of the geometry LL and the length scale required to develop the acoustic nature, i.e. the thickness of the wavefront of the pressure wave.

Equation (12) also provides a theoretical insight into the fact that StSt reflects the fluid compressibility, as reported by Reijers et al. [36]. For St≪1St\ll1, equation (12) can be rewritten as

∂2Φ∗∂x∗2=0,(14)\frac{\partial^2\Phi^*}{\partial x^{*2}}=0, \tag*{(14)}

which describes incompressible and irrotational flow, indicating that the pressure fluctuations due to wave propagation are negligible, as experimentally evidenced in Figure 1. In contrast, for St≫1St\gg1, the liquid compressibility and pressure fluctuations become visible (Figure 1). The pressure fluctuations are derived from the Euler equation (10). Integrating this equation with respect to xx and assuming p=p0p=p_0 and u=0u=0 in the far field gives

p−p0=ρcU,(15)p-p_0=\rho cU, \tag*{(15)}

where ρcU\rho cU is known as the water hammer pressure; this expression is known as Joukovski’s equation [5, 37], a classical formulation for predicting the maximal pressure fluctuations in one-dimensional flows. As explained earlier, in the limit Δt≈0\Delta t\approx0, an instantaneous pressure rise is assumed and the pressure in the liquid jumps from zero to ρcU\rho cU. According to the above discussion, in our experiments where the characteristic speed is U=u0U=u_0, the development of pressure P‾\overline{P} is expected to be dominated by the Strouhal number StSt, which successively connects P‾=0\overline{P}=0 at St≪1St\ll1 and P‾=ρcu0\overline{P}=\rho c u_0 at St≫1St\gg1.

Experimental results and remarks

Pressure fluctuations in liquids

Figure 6 compares the amplitude of the spatially averaged liquid pressure change (hereafter, the pressure fluctuation) P‾\overline{P} to the Strouhal number StSt for different values of Δt\Delta t and LL. The use of different floor materials results in Δt\Delta t varying from 0.1–2.2 ms, as indicated in the legend. The use of water (△\triangle), 1 cSt silicone oil (▽\triangledown), and ethanol (⋄\diamond) allows us to vary the speed of sound cc and the liquid density ρ\rho, which are related to the water hammer pressure. In particular, the speed of sound cc in water is approximately 1.6 times faster than that in 1 cSt silicone oil. The density ρ\rho of water is approximately 1.3 times higher than that of ethanol.

Image not extracted

Figure 6. Magnitude of pressure fluctuation P‾/(ρcu0)\overline{P}/(\rho c u_0) versus Strouhal number StSt. Data points show experimental results and each plot colour denotes a different floor type. The shapes of the data points correspond to the different liquid types (including a hydrogel), as indicated in the legend.

The experimental data fall onto a single curve that is insensitive to the experimental parameters tested. All of the data for St>0.2St > 0.2 overlap significantly and approach P‾/(ρcu0)≈1.0\overline{P}/(\rho c u_0) \approx1.0 when St>O(1)St > O(1), whereas the data using resin floors for St≤0.2St \le0.2 are scattered. The collapse of the data in Figure 6 indicates that the Strouhal number StSt describes the gradual pressure development in a one-dimensional tube. Of the conditions tested, St∼0.2St \sim0.2 is the threshold at which the pressure fluctuations become visible and the transition begins (P‾/(ρcu0)≈0.1\overline{P}/(\rho c u_0) \approx0.1).

Figure 6 also contains data for 10 cSt silicone oil (◯\bigcirc). This has a greater viscosity than the other liquids, but the same trend can be observed because the viscous boundary layer δ∼νΔt∼O(10−2)\delta\sim\sqrt{\nu\Delta t} \sim O(10^{-2}) mm is expected to be sufficiently thinner than the container radius ∼O(1)\sim O(1) mm, as argued in the case of jetting experiments ([33, 16]). The surface tension does not significantly affect the pressure fluctuations in the present experiments, despite the surface tension of water being approximately 4.3 times higher than that of 1 cSt silicone oil. From the above, the Strouhal number StSt provides a powerful means of describing the pressure fluctuations inside widely used low-viscosity liquids.

Previous research on similar systems suggests that the accelerating fluid can be modelled as either a compressible fluid ([12, 25, 10, 41, 18]) or an incompressible fluid ([31, 23, 9, 14, 34, 16, 19, 26]), even if the Strouhal number is in the intermediate region St≈O(1)St \approx O(1). Our experimental results (Figure 6) obviously show that the pressure wave appears at moderate values of StSt and its magnitude scales with StSt: the transition from the “incompressible fluid” region starts at St∼O(10−1)St \sim O(10^{-1}) (St≈0.2St \approx0.2 in this experiment) and the “compressible fluid” region is established at St∼O(101)St \sim O(10^{1}). This is reasonably consistent with the case of laser-ablation onto a droplet ([36]), in which fluid compressibility appears in the flow field (and accordingly the pressure field) at St≫1St \gg1.

Cavitation inception following pressure fluctuations

It is thus expected that the Strouhal number StSt could provide a measure of the tendency for cavitation in a one-dimensional fluid system. The cavitation number for the accelerated liquid ([34]) is written as

Ca=Patm−PvρaL,(16)Ca = \frac{P_{atm} - P_v}{\rho a L}, \tag*{(16)}

where a=U‾/Δta = \overline{U}/\Delta t. Employing the first-order approximation ρL(U‾/Δt)∼Stρcu0\rho L(\overline{U}/\Delta t) \sim St \rho c u_0, we obtain

Ca∼St−1Patm−PvρcU‾.(17)Ca \sim St^{-1} \frac{P_{atm} - P_v}{\rho c \overline{U}}. \tag*{(17)}

This expression makes sense for situations in which the influence of Δt\Delta t is dominant over that of other parameters. If Δt\Delta t is dominant, CaCa and St−1St^{-1} play a similar role and are thus interchangeable. Increasing StSt decreases CaCa, meaning a greater pressure reduction, as expected, and a higher possibility of cavitation occurring [9, 34, 11]. This perhaps explains the findings in previous studies [14, 34, 38] (Xu et al. 2021), in which the overall cavitation tendency was reasonably classified by only considering CaCa as the significance of Δt\Delta t remained largely unchanged.

Figure 7 shows the probability curves of cavitation in silicone oil (10 cSt) for different floor types. Each marker shows the cavitation probability determined through 10 separate experimental runs. The gradual increase in probability is fitted by the sigmoid function [30, 17, 8, 32, 7] as

Prob.=12[1+erf⁡(St−αβ2)],(18)Prob. = \frac{1}{2}\left[1+\operatorname{erf}\left(\frac{St-\alpha}{\beta\sqrt{2}}\right)\right], \tag*{(18)}

where erf⁡()\operatorname{erf}() is the error function implemented in Matlab, and α\alpha and β\beta are the fitting parameters. Overall, cavitation starts at StSt values as low as 0.2, which agrees well with the condition for noticeable pressure fluctuations. Closer observation reveals the influence of the floor material. For relatively stiff floors (i.e. SS400 and aluminium), a 50% probability of cavitation is reached at St=0.43St = 0.43 and St=0.46St = 0.46, respectively. For softer floors, the probability curve is steeper (i.e. smaller β\beta values), and the minimum StSt value required does not vary significantly. The physical reasoning behind this remains unclear, and further investigation is needed. The Δt\Delta t value varies for softer floors (Figure 6), so its influence on the results needs to be clarified.

Pressure fluctuations and cavitation in a hydrogel

The Strouhal number StSt is still a powerful tool for describing the pressure fluctuations in hydrogels. In Figure 6, the gelatin data (squares) for various floor types are in good agreement with the other data for standard fluids. The 5 wt% gelatin gel is a water-based viscoelastic material, which behaves as a soft solid in the rest state. At the moment of impact, it is expected that the gelatin gel could flow due to the fast deformation (Δt∼O(0.1)\Delta t \sim O(0.1) ms). The above discussion on the Strouhal number StSt and the water hammer pressure ρcu0\rho c u_0 can be applied to the accelerated hydrogel, although further investigations are needed for a quantitative understanding. Our findings provide an experimental understanding of the inception of cavitation inside the gel ((Kang et al. 2017a,b; Kang & Raphael 2018)); see the supplementary materials. This phenomenon has been used to address brain injury issues ([27] (Yu et al. 2020)).

Analysis of the pressure fluctuations

We now discuss the basic theoretical framework for predicting the pressure fluctuations. First, we consider a gradual pressure increment at the pressure wavefront to confirm the smooth connection between the “incompressible fluid” region (St≪1St \ll1) and the “compressible fluid” region (St>1St > 1) as a function of the Strouhal number StSt (§ 3.3.1). We then compare the theoretical models and the experimental data (§ 3.3.2). We discuss the influence of both the container mass and the pressure wavefront profile.

Model considering thickness of pressure wave

This paper focuses on the case of St=L/(cΔt)≈1St = L/(c\Delta t) \approx1, where the length of the liquid column LL is similar to the length cΔtc\Delta t. We call cΔtc\Delta t the thickness of the pressure wavefront because it corresponds to the length required to establish the water hammer pressure. In such cases, the pressure fluctuations are not determined by the classical formulation assuming an instantaneous pressure rise [equation (3.7)], but are caused by both the temporal pressure change during Δt\Delta t and the reflection of the pressure wave at the boundaries.

The liquid column in the container has a gas–liquid interface at one end and a solid–liquid interface at the other end. First, we define the function f∗(ξ∗,t∗)f^{*}(\xi^{*},t^{*}) describing the liquid pressure change, where ξ∗\xi^{*} and t∗t^{*} are the distance from the solid–liquid interface and the time from the initial impact, respectively. The pressure change propagates in the direction of ξ∗\xi^{*}. The wavefront is located at ξ∗=t∗\xi^{*}=t^{*} at t∗t^{*}. We express f∗(ξ∗,t∗)f^{*}(\xi^{*},t^{*}) as

f∗(ξ∗,t∗)={1(t∗−1≥ξ∗)I∗(ξ∗−t∗+1)(t∗−1≤ξ∗≤t∗)0(t∗≤ξ∗),(19)f^{*}(\xi^{*},t^{*}) = \begin{cases} 1 & (t^{*}-1 \ge\xi^{*})\\ I^{*}(\xi^{*}-t^{*}+1) & (t^{*}-1 \le\xi^{*} \le t^{*})\\ 0 & (t^{*} \le\xi^{*}), \end{cases} \tag*{(19)}

where I∗(η)I^{*}(\eta) is the pressure change in the pressure wavefront, which is modelled as

I∗(η)=12sin⁡(ηπ+π2)+12;(20)I^{*}(\eta)=\frac{1}{2}\sin\left(\eta\pi+\frac{\pi}{2}\right)+\frac{1}{2}; \tag*{(20)}

see also Figure 8. The reflection of the pressure wave at the boundaries is modelled by assuming the principle of superposition [37]. We assume that the mismatching of the acoustic impedance at the gas–liquid and solid–liquid interfaces is significant, and hence we neglect the energy loss upon reflection. The sign of the pressure reverses when the pressure wave reflects at the gas–liquid interface, but does not change when the pressure wave collides the solid–liquid interface. The local pressure in the liquid column p∗p^{*} is calculated as follows:

Image not extracted

Figure 8. Outline view of the function of pressure change f∗f^*, described by the dimensionless time t∗t^* and dimensionless position ξ∗\xi^*. Red, blue, and black lines show models with the assumption that the thickness of the wavefront is 0 (i.e. water hammer theory), that the liquid is incompressible, and that the wavefront has a finite thickness, as modelled herein, respectively.

p∗(ξ∗,t∗,St)=∑i=1ceil⁡(t∗St)g∗(ξ∗,t∗,St,i),(21)p^{*}(\xi^{*},t^{*},St)=\sum_{i=1}^{\operatorname{ceil}\left(\frac{t^{*}}{St}\right)}g^{*}(\xi^{*},t^{*},St,i), \tag*{(21)}
g∗(ξ∗,t∗,St,k)={f∗(kSt−ξ∗,t∗)(k=4n)−f∗((k−1)St+ξ∗,t∗)(k=4n−1)−f∗(kSt−ξ∗,t∗)(k=4n−2)f∗((k−1)St+ξ∗,t∗)(k=4n−3),g^{*}(\xi^{*},t^{*},St,k)= \begin{cases} f^{*}(kSt-\xi^{*},t^{*}) & (k=4n)\\ -f^{*}((k-1)St+\xi^{*},t^{*}) & (k=4n-1)\\ -f^{*}(kSt-\xi^{*},t^{*}) & (k=4n-2)\\ f^{*}((k-1)St+\xi^{*},t^{*}) & (k=4n-3), \end{cases}

where nn is a natural number.

This model allows us to estimate the temporal changes in pressure. We then calculate the amplitude of the spatially averaged pressure in the liquid, P‾\overline{P}. Hereafter, the amplitude of the pressure change P‾/(ρcu0)\overline{P}/(\rho c u_{0}) calculated from the above model is compared with that estimated from the experiments, mV/(ρALu0)mV/(\rho A L u_{0}).

Comparison and discussion

We now compare the experimental results (dataset visualised in Figure 6) with the model in Figure 9(a). The black solid line expresses our model with consideration of the pressure wavefront thickness (see § 3.3.1). Figure 9(b) shows a magnified view. For St=O(10−2)St=O(10^{-2}), the fluctuations are not obvious in either the model or the experimental results (P‾/(ρcu0)<10−2\overline{P}/(\rho c u_{0})<10^{-2}), suggesting that the liquid behaves as an incompressible fluid. For St=O(10−1)St=O(10^{-1}), the fluctuations in both the experimental results and the model increase notably, which is associated with an increase in the Strouhal number StSt. Therefore, the effect of liquid compressibility appears at St=O(10−1)St = O(10^{-1}). The proposed model predicts the onset of pressure fluctuations (see Figure 9b, St∼0.2St \sim0.2) and captures the overall trend of these fluctuations with respect to StSt, but overestimates the magnitude of the fluctuations for higher StSt values (St<O(100)St < O(10^{0})). Though the model implies that the water hammer pressure (P‾/(ρcu0)=1.0\overline{P}/(\rho c u_0) = 1.0) is reached at higher StSt values, the experimental results saturate at around P‾/(ρcu0)∼0.7\overline{P}/(\rho c u_0) \sim0.7 for St∼1St \sim1. Although no data could be obtained for St>O(101)St > O(10^{1}) due to experimental limitations, we expect that the experimental scenario would differ for higher StSt values as well because of the influence of the container motion and the profile for the pressure wavefront, as discussed below.

Image not extracted

Figure 9. Dimensionless amplitude of the spatial pressure in a liquid as a function of the Strouhal number, StSt. Data points show experimental results and each plot colour corresponds to a different floor type. The shapes of the data points correspond to different liquids. The red and blue lines show the pressure calculated under the assumptions compressible and incompressible liquids, respectively. The black thin line represents the calculation result using the model (§ 3.3.1).

First, we investigate the effects of the container motion. The momentum conservation law for a system consisting of a container and a liquid is

mv+Aρ∫0Lu dx=H(const.),(22)mv + A\rho\int_{0}^{L} u\,\mathrm{d}x = H(\mathrm{const.}), \tag*{(22)}

where mm is the mass of the container, vv is the velocity of the container, and AA is the cross-sectional area of the liquid column. Note that we neglect the contribution of gravitational acceleration because it is expected to be very small when compared with the acceleration imposed by the impact (g≪u0/Δtg \ll u_0/\Delta t). Equation (22) can be reformulated as follows in dimensionless quantities:

mv∗u0+Aρu0cΔt∫0Stu∗ dx∗=H(const.).(23)mv^*u_0 + A\rho u_0 c\Delta t \int_{0}^{St} u^*\,\mathrm{d}x^* = H(\mathrm{const.}). \tag*{(23)}

The momentum conserved in the system, HH, is calculated using v∗v^* and u∗u^* immediately after the impact. The velocity of the container v∗v^* is then calculated from the constant momentum HH and the instantaneous velocity in the liquid u∗u^*. We incorporate the function f∗f^* in § 3.3.1 into the velocity of the container, v∗v^*, as

f∗(ξ∗,t∗)={v∗(t∗−1≥ξ∗)Γ∗(ξ∗−t∗+1)(t∗−1≤ξ∗≤t∗)0(t∗≤ξ∗).(24)f^*(\xi^*,t^*) = \begin{cases} v^* & (t^*-1 \ge\xi^*)\\ \Gamma^*(\xi^*-t^*+1) & (t^*-1 \le\xi^* \le t^*)\\ 0 & (t^* \le\xi^*). \end{cases} \tag*{(24)}

The pressure in the liquid is calculated from equation (21) using the function f∗f^*. The velocity of the liquid, u∗u^*, is calculated from the following equation:

u∗(ξ∗,t∗,St)=∑i=1ceil⁡(t∗St)h∗(ξ∗,t∗,St,i),(25)u^*(\xi^*,t^*,St)=\sum_{i=1}^{\operatorname{ceil}\left(\frac{t^*}{St}\right)}h^*(\xi^*,t^*,St,i), \tag*{(25)}
h∗(ξ∗,t∗,St,k)={−f∗(kSt−ξ∗,t∗)(k=4n)−f∗((k−1)St+ξ∗,t∗)(k=4n−1)f∗(kSt−ξ∗,t∗)(k=4n−2)f∗((k−1)St+ξ∗,t∗)(k=4n−3).h^*(\xi^*,t^*,St,k)= \begin{cases} -f^*(kSt-\xi^*,t^*) & (k=4n)\\ -f^*((k-1)St+\xi^*,t^*) & (k=4n-1)\\ f^*(kSt-\xi^*,t^*) & (k=4n-2)\\ f^*((k-1)St+\xi^*,t^*) & (k=4n-3). \end{cases}

The velocity of the container, v∗v^*, is dependent on the momentum of the liquid, which implies some dependence on LL, mm, AA, and ρ\rho (23), and can be interchanged with the pressure, as we did for the experimental data. Considering the velocity change of the container, the amplitude of the pressure fluctuations in the liquid is not only defined by the Strouhal number StSt and the water hammer pressure.

The revised model is compared with the experimental results in the case of an epoxy resin floor (Figure 10). We used two different tubes to vary the mass; the filled and open circles represent the data for the long and short tubes, respectively. The dashed and dotted lines correspond to the model for the long and short tubes, respectively. The revised model taking the change in container velocity into account improves accuracy in terms of the overestimation observed in the original models. The revised model for different tubes also predicts the influence of the tube types, whereas such variations might be not necessarily visible in the experimental data.

Effect of container motion on the pressure fluctuations in a liquid

Figure 10. Effect of container motion on the pressure fluctuations in a liquid. Open (filled) circles show experimental results using short (long) tube. The dotted (dashed) line shows the model results considering the motion of the short (long) container.

We also tested the effect of the pressure wavefront profile. For comparison purposes, we used a simple profile, illustrated by the dot-dashed line in Figure 11(a). This is expressed by I∗(η)I^{*}(\eta) in equation (19), which is defined as

Profile of pressure wavefront and dimensionless fluctuations in water hammer pressure

Figure 11. (a) Profile of pressure wavefront. (b) Dimensionless fluctuations in water hammer pressure with respect to Strouhal number. The solid and dot-dashed lines represent different wavefront profiles.

I∗(η)=1−η.(26)I^{*}(\eta) = 1 - \eta. \tag*{(26)}

This model overpredicts the pressure fluctuations for smaller StSt values (St<O(10−1)St < O(10^{-1})), but gives better agreement for larger StSt values (Figure 11b). This suggests that the profile of the pressure wavefront could affect the accuracy of the model predictions, which provides a possible explanation for the difference between the experimental results and the original model (Figure 9). In our experiments, the profile of the pressure wavefront is assumed to be dependent on the floor material. Hence, the effect of the pressure profile causes variations in the pressure fluctuations measured experimentally at the same StSt as well as the cavitation trend for each floor type.

In addition to the above considerations, there are two possible reasons for differences between the model and experiments: the vibration of the container and the curvature of the bottom of the container. Although the model does not consider the effect of vibrations in the container, this phenomenon was indeed observed in the experiments (see Figure 4a). The momentum in the experimental system is converted to vibrations. The vibration of the container after the rebound might also be related to the amount of liquid inside the system ([23, 25, 1]). Additionally, the model does not consider the curvature of the bottom of the container, which leads to the formation of a plane wave in this region. However, we used a container with a rounded bottom in the experiments. A radial phase lag is likely to occur and the fluctuations will differ from those of the model.

Conclusions

This paper has shown that the Strouhal number StSt provides a measure of the fluid compressibility in a one-dimensional system, unlike other dimensionless indicators such as MaMa and FrFr. Experiments with various liquid column depths inside a tube were performed at different liquid velocities, acceleration durations, and media types (four standard liquids and one hydrogel), producing a wide range of Strouhal numbers, 0.02≤St≤2.20.02 \le St \le2.2. The pressure fluctuations, which were calculated based on the measured acceleration, indicate a unified trend as a function of StSt for a wide range of parameters (Figure 6), suggesting that StSt gives a good measure of the liquid compressibility in cases where the pressure impulse approach breaks down. Our experimental findings regarding the importance of StSt are consistent with those from numerical work on bursting droplets [36], suggesting that our discussion related to StSt might even be applicable in different geometries. We also found that StSt is useful for describing the cavitation tendency (Figure 7). In addition to empirical relations, we derived a conceptual model in which a modified version of water hammer theory accounts for the finite thickness of the pressure wavefront. This simple model describes the overall trend of the experimental data: the magnitude of the pressure fluctuations increases as the Strouhal number StSt increases (Figure 9). Implementing corrections in the model allowed us to show that both the motion of the surrounding container and the profile of the pressure wavefront influence the pressure development (see Figure 10 and Figure 11).

Image not extracted

Figure 7. Cavitation probability of silicone oil (10 cSt) for various floor materials, as indicated by colours. The probabilities were experimentally determined through 10 separate experimental runs. The fitting curves were obtained based on equation (3.10), where the fitting parameters were α=0.431\alpha= 0.431 and β=0.109\beta= 0.109 for SS400, α=0.456\alpha= 0.456 and β=0.113\beta= 0.113 for aluminium, α=0.396\alpha= 0.396 and β=0.053\beta= 0.053 for epoxy resin, and α=0.249\alpha= 0.249 and β=0.0032\beta= 0.0032 for ABS resin.

Acknowledgements

We thank Prof. J. Rodríguez-Rodríguez for fruitful discussions. We also thank Dr. Andres Franco-Gomez, Prof. Masakazu Muto, Dr. Hajime Onuki, Dr. Prasad Sonar from Tokyo University of Agriculture and Technology for proofreading the manuscript and making a number of helpful suggestions. This work was partly supported by JSPS KAKENHI through grant numbers JP16J08521, JP17H01246, JP20H00223, JP20H00222, JP20K20972, and JP23K19089, the Japan Science and Technology Agency PRESTO (Grant No. JPMJPR21O5), and Japan Agency for Medical Research and Development (Grant No. JP22he0422016) and by funding from the Institute of Global Innovation Research at the Tokyo University of Agriculture and Technology.

Supplementary materials

Cavitation in a hydrogel system

The onset of cavitation and the associated dynamics are of great interest in the community, as they contribute to understanding the brain injuries that may occur following a sudden impact [4, 29]. Though a similar system observed the onset of cavitation in a gelatin hydrogel using a high-speed camera [34], the present study did not perform any direct visualizations. However, comparing acceleration data as a function of time, some cases exhibit enhanced signals at O(1)O(1) ms after the impact (see Figure 12). These might be related to the perturbation inside the gelatin column, namely the onset of cavitation. This result supports our claim that the present research has the potential to clarify our understanding of acceleration-induced brain injuries, although the sensitivity of measurements is limited.

Acceleration $a$ as a function of time $t$ with and without cavitation

Figure 12. Acceleration aa (m/s2^{2}) measured as a function of time tt (s). The orange line represents the data with cavitation, while the blue line shows data without cavitation onset. The liquid column height and drop height are L=60L = 60 mm and H=15H = 15 mm. The Strouhal number StSt is expected to be greater than 0.2. The signals have not been low-pass filtered to avoid the elimination of high-frequency responses caused by bubble activity.

References

  1. [1]Andrade, K., Catalán, J., Marín, J. F., Salinas, V., Castillo, G., Gordillo, L. & Gutiérrez, P. 2023 Swirling fluid reduces the bounce of partially filled containers. Phys. Rev. Lett. 130, 244001.DOI
  2. [2]Antkowiak, A., Bremond, N., Le Dizès, S. & Villermaux, E. 2007 Short-term dynamics of a density interface following an impact. J. Fluid Mech. 577, 241–250.DOI
  3. [3]Bao, H., Reuter, F., Zhang, H., Lu, J. & Ohl, C. D. 2023 Impact-driven cavitation bubble dynamics. Experiments in Fluids 64 (27).DOI
  4. [4]Barney, C. W., Dougan, C. E., McLeod, K. R., Kazemi-Moridani, A., Zheng, Y., Ye, Z., Tiwari, S., Sacligil, I., Riggleman, R. A., Cai, S., Lee, J. H., Peyton, S. R., Tew, G. N. & Crosby, A. J. 2020 Cavitation in soft matter. Proc. Natl. Acad. Sci. 117 (17), 9157–9165.DOI
  5. [5]Batchelor, G. K. 1967 An Introduction to Fluid Dynamics. Cambridge University Press.
  6. [6]Bergant, A., Simpson, A. R. & Tisseling, A. S. 2006 Water hammer with column separation: A historical review. J. Fluids Struct. 22 (2), 135–171.DOI
  7. [7]Bustamante, M. C. & Cronin, D. S. 2019 Cavitation threshold evaluation of porcine cerebrospinal fluid using a polymeric split Hopkinson pressure bar-confinement chamber apparatus. J. Mech. Behav. Biomed. Mater. 100, 103400–103409.DOI
  8. [8]Bustamante, M. C., Singh, D. & Cronin, D. S. 2017 Polymeric opkinson bar-confinement chamber apparatus to evaluate fluid cavitation. Exp. Mech. 58, 55–74.DOI
  9. [9]Daily, J., Pendlebury, J., Langley, K., Hurd, R., Thomson, S. & Truscott, T. 2014 Catastrophic cracking courtesy of quiescent cavitation. Phys. Fluids 26 (9), 091107.arxiv.org/abs/1210.3764
  10. [10]Daou, M. M., Igulada, E., Dutilleul, H., Citerne, J.-M., Rodríguez-Rodríguez, J., Zaleski, S. & Fuster, D. 2017 Investigation of the collapse of bubbles after the impact of a piston on a liquid free surface. AIChE J. 63 (6), 2483–2495.DOI
  11. [11]Eshraghi, J., Veilleux, J.-C., Shi, G., Collins, D., Ardekani, A. M. & Vlachos, P. P. 2022 Assessment of cavitation intensity in accelerating syringes of spring-driven autoinjectors. Pharmaceutical Research 39, 2247–2261.DOI
  12. [12]Fogg, D. W. & Goodson, K. E. 2009 Bubble-induced water hammer and cavitation in microchannel flow boiling. J. Heat Transf. 131 (12), 121006.DOI
  13. [13]Fujikawa, S., Takeda, Y., Yano, Y. & Murai, Y. 2005 Foundation of Engineering Fluid Mechanics. Baifukan (in Japanese).
  14. [14]Garcia-Atance Fatjo, G. 2016 New dimensionless number to predict cavitation in accelerated fluid. Int. J. Comput. Methods Exp. Meas. 4 (4), 484–492.DOI
  15. [15]Ghidaoui, M. S., Zhao, M., McInnis, D. A. & Axworthy, D. H. 2005 A review of water hammer theory and practice. Applied Mechanics Reviews 58, 49–76.DOI
  16. [16]Gordillo, J. M., Onuki, H. & Tagawa, Y. 2020 Impulsive generation of jets by flow focusing. Journal of Fluid Mechanics 894, A3.DOI
  17. [17]Hayasaka, K., Kiyama, A. & Tagawa, Y. 2017 Effects of pressure impulse and peak pressure of a shockwave on microjet velocity in a microchannel. Microfluid. Nanofluid. 21 (11), 166.arxiv.org/abs/1703.04843
  18. [18]Kamamoto, K., Kiyama, A., Tagawa, Y. & Zhang, X. 2021a Ouzo column under impact: Formation of emulsion jet and oil-lubricated droplet. Langmuir 37 (6), 2056–2064, pMID: 33527827.DOI
  19. [19]Kamamoto, K., Onuki, H. & Tagawa, Y. 2021b Drop-on-demand painting of highly viscous liquids. Flow 1, E6.DOI
  20. [23]Killian, T. W., Klaus, R. A. & Truscott, T. T. 2012 Rebound and jet formation of a fluid-filled sphere. Physics of Fluids 24 (12), 122106.DOI
  21. [24]Kiyama, A., Mansoor, M. M., Speirs, N. B., Tagawa, Y. & Truscott, T. T. 2019 Gelatine cavity dynamics of high-speed sphere impact. J. Fluid Mech. 880, 707–722.DOI
  22. [25]Kiyama, A., Tagawa, Y., Ando, K. & Kameda, M. 2016 Effects of water hammer and cavitation on jet formation in a test tube. J. Fluid Mech. 787, 224–236.DOI
  23. [26]Krishnan, S., Bharadwaj, S. V. & Vasan, V. 2022 Impact of freely falling liquid containers and subsequent jetting. Experiments in Fluids 63, 108.DOI
  24. [27]Lang, J., Nathan, R., Zhou, D., Zhang, X., Li, B. & Wu, Q. 2021 Cavitation causes brain injury. Phys. Fluids 33, 031908.DOI
  25. [28]Leighton, T.G. 1994 The Acoustic Bubble. Academic Press.
  26. [29]Marsh, J. L. & Bentil, S. A. 2021 Cerebrospinal fluid cavitation as a mechanism of blast-induced traumatic brain injury: A review of current debates, methods, and findings. Front. Neurol. 12, 626393.DOI
  27. [30]Maxwell, A. D., Cain, C. A., Hall, T. L., Fowlkes, J. B. & Xu, Z. 2013 Probability of cavitation for single ultrasound pulses applied to tissues and tissue-mimicking materials. Ultrasound Med. Biol. 39 (3), 449–465.DOI
  28. [31]Niederhaus, C. E. & Jacobs, J. W. 2003 Experimental study of the richtmyer–meshkov instability of incompressible fluids. Journal of Fluid Mechanics 485, 243–277.DOI
  29. [32]Oguri, R. & Ando, K. 2018 Cavitation bubble nucleation induced by shock-bubble interaction in a gelatin gel. Phys. Fluids 30 (5), 051904.DOI
  30. [33]Onuki, H., Oi, Y. & Tagawa, Y. 2018 Microjet generator for highly viscous fluids. Phys. Rev. Appl. 9 (1), 014035.DOI
  31. [34]Pan, Z., Kiyama, A., Tagawa, Y., Daily, J. L., Thomson, S., Hurd, R. & Truscott, T. T. 2017 Cavitation onset caused by acceleration. Proc. Natl. Acad. Sci. 114 (32), 8470–8474.DOI
  32. [35]Rapet, J., Tagawa, Y. & Ohl, C. D. 2019 Shear-wave generation from cavitation in soft solids. Appl. Phys. Lett. 114 (12), 123702–123705.DOI
  33. [36]Reijers, S. A., Snoeijer, J. H. & Gelderblom, H. 2017 Droplet deformation by short laser-induced pressure pulses. J. Fluid Mech. 828, 374–394.DOI
  34. [37]Thompson, P. A. 1972 Compressible Fluid Dynamics. McGraw-Hill Inc.
  35. [38]Wang, Z., Liu, S., Li, B., Zuo, Z. & Pan, Z. 2022 Large cavitation bubbles in the tube with a conical-frustum shaped closed end during a transient process. Physics of Fluids 34 (6), 063312.DOI
  36. [41]Yukisada, R., Kiyama, A., Zhang, X. & Tagawa, Y. 2018 Enhancement of focused liquid jets by surface bubbles. Langmuir 34 (14), 4234–4240.DOI

Paper details

vlnc:W_8cz4spsa
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 6

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents