Cavitation Intensity Mechanism in the Hydrodynamic Cavitation Abrasive Finishing Open Access

Internal surface machining is a critical step in manufacturing channel components, ensuring surface quality and cleanliness after complex design and fabrication [1]. Laser powder bed fusion (LPBF) has demonstrated tremendous capability in designing and manufacturing parts containing fluid channels for reducing weight, and improving flow resistance and thermal properties [2,3]. However, LPBF fabricated channels inevitably exhibit issues such as non-uniform material distribution, poor surface quality in the overhanging regions, and slag and powder adhesion on the inner walls [4]. In mechanical and aerospace applications, higher inner surface quality is often required to minimize fluid pressure loss and particle shedding. Currently, inner surface processing methods for channels, including electrochemical polishing [5], chemical polishing [6], and abrasive flow polishing [7,8], face challenges such as uneven inner treatment, excessive material removal, and environmental pollution, particularly in the LPBF channels. In contrast, the hydrodynamic cavitation abrasive finishing (HCAF), a novel non-contact, flexible, and clean fluid polishing technology, effectively mitigates issues such as abrasive agglomeration and material loss [9]. When processing highly uneven inner surfaces in additive manufacturing (AM), it can improve the surface finish by 90% [10].

The HCAF builds upon traditional cavitation jet surface finishing by introducing micron-sized abrasive particles (e.g., SiC, Al₂O₃, or CeO₂). These particles interact synergistically with cavitation bubbles, enhancing the erosion effect to facilitate the material removal, thereby achieving finer and more efficient surface polishing. Nagalingam and Yeo first proposed the use of an HCAF method and developed the first cavitation jet system [9]. In this system, abrasives first mix uniformly with water in a high-pressure pipeline to form an abrasive slurry, which then passes through the cavitation region to generate a jet, known as a pre-mixed hydrodynamic cavitation abrasive finishing. This method ensures good mixing performance but involves a complex setup and causes wear on the high-pressure pump and pipelines. Using this pre-mixed jet machining approach, the inner surface of AlSi10Mg flow channels was polished. Nagalingam et al. revealed the synergistic mechanism between cavitation erosion and abrasives and systematically investigated the effects of pressure, flow velocity, and temperature on the surface roughness [11]. Subsequently, a multi-jet hydrodynamic cavitation abrasive finishing (MJ-HCAF) system was developed and applied to the surface finishing of complex internal channels in additive manufacturing [10]. Compared to the previous pre-mixed system, the new post-mixed HCAF device delivers abrasives and water separately, mixing them at or after the cavitation region. This design effectively prevents abrasive-induced damage to the high-pressure pump and pipelines while enhancing the durability and maintenance convenience of the equipment. A summary of relevant literature on the HCAF of AM parts is presented in Table 1.

Therefore, to extend the system lifespan, this study employs a post-mixed MJ-HCAF system. It is well known that in HCAF, which is based on the synergistic interaction between the cavitation bubbles and abrasives, the cavitation intensity significantly influences machining efficiency. Previous studies have shown that the cavitation intensity of water jets is influenced by the concentration of cavitation bubbles, and modifying nozzle structures and hydrodynamic parameters can enhance the cavitation intensity [15,16]. Wu et al. [17] used a high-speed camera to capture the growth and development of cavitation clouds, revealing that the nozzle throat cross-sectional size has the most significant impact on the cavitation effect. Additionally, Soyama [18] found that in non-submerged cavitation water jets, the velocity of the external low-speed water jet also significantly affects the cavitation intensity. However, little research has been conducted on the evolution of cavitation bubbles in the high-speed jet region, including their generation, growth, collapse, and influencing factors within the mixing chamber and processing pipeline.

This study focuses on an in-depth investigation of the cavitation jet mechanism in the HCAF. The study examines the effects of various parameters, including abrasive branch-end pressure, nozzle size, and channel dimensions, on the cavitation intensity during the HCAF processing of channels. Additionally, it explores their influence on the evolution of cavitation bubbles. Furthermore, a high-speed camera was used to record the cavitation cloud intensity under different parameters. Destructive measurement methods were employed to document the surface morphology after internal channel polishing, followed by analysis and interpretation of the results. The findings provide guidance for the optimization of the HCAF mixed-flow chamber design, as well as the selection of polishing channel dimensions and operating pressure.

Based on the hydrodynamic cavitation abrasive finishing mechanism, a hydrodynamic cavitation abrasive finishing system was constructed. Figure 1 illustrates the overall structure of the hydrodynamic cavitation abrasive finishing system, which consists of a high-pressure cavitation circuit, a low-pressure abrasive circuit, and a mixed-flow chamber that connects both circuits. Figure 2 presents a photograph of the actual HCAF system. The primary function of the high-pressure cavitation circuit is to generate cavitation jets. It comprises a water tank, a pre-filter, a high-pressure plunger pump, an accumulator, a pressure regulating valve, and a mechanical pressure gauge, all integrated within the control cabinet. The high-pressure plunger pump is a three-piston pump with a maximum pressure of 25 MPa and a maximum flowrate of 15 l/min. The primary function of the low-pressure abrasive circuit is to transport abrasives. It consists of a carrier water tank, a ball valve, a filtration water tank, a precision filter, an electric diaphragm pump, and a pressure gauge.

The mixed-flow chamber is a key component of the system, designed to merge and homogenize the high-pressure cavitation jet and the low-pressure abrasive flow, forming an abrasive cavitation jet for cavitation-based polishing of the workpiece connected downstream, as shown in Fig. 3(a). Additionally, the diagram of the mechanism of cavitation and abrasive synergistic action is shown in Fig. 4.

The mixed-flow plate is the core component of the mixed-flow chamber, responsible for blending the high-pressure cavitation jet and the low-pressure abrasive flow. In this study, it was optimized and improved. The mixed-flow plate was designed with a “T”-shaped mixing structure, modified from the “V”-shaped structure originally developed by Nagalingam et al. Compared to the V-shaped mixing structure, which is prone to backflow, the improved T-shaped design features a self-priming negative pressure function. The specific dimensions are shown in Fig. 3(b).

As shown in Fig. 1, the low-pressure abrasive flow loop uses a 25-watt low-flow electric diaphragm pump to deliver the abrasive flow to the mixing chamber. The pump is driven by a stepper motor and allows step-less speed adjustment to achieve variable flowrates and pressures. Specifically, the abrasive flow pressure of –0.04 MPa (P_u2) is generated by the negative pressure structure of the mixing plate during the operation of the 15 MPa cavitation jet. The pressure range of –0.04 to 0.06 MPa is achieved through step-less speed control of the abrasive flow pump. The abrasive particles used in the experiment had an average size of approximately 60 μm. The scanning electron microscope (SEM) image of the abrasives is shown in Fig. 5.

The mixed-flow chamber is connected to the polishing specimen at the bottom. In this study, the workpiece used was a 316L channel with a diameter of 1.8 mm, a length of 100 mm, and a wall thickness of 2.1 mm, as shown in Fig. 3(c). The LPBF machine used in this study was the Zhongrui iSLM160. The feedstock was 316L spherical powder with a loose packing density of 3.9 g/cm³ and a particle size range of 15–53 μm.

A high-speed camera system was used to capture the distribution of cavitation bubbles inside the flow channel. The high-speed camera used was the FASTCAM Mini AX200, capable of recording 20,000 frames per second at a resolution of 640 × 480. Shooting parameters and image recording were adjusted via dedicated software on a computer. The flow channel was fabricated using photopolymerization-based 3D printing with a transparent resin material (ZR820) that has 85% light transmittance. Dimensional fine-tuning ensured that the printed flow channel closely matched the diameter and length of the metal additive-manufactured counterpart, maintaining similar cavitation bubble distribution characteristics.

A confocal microscope was used to capture the surface morphology of the channel and measure its surface roughness. The confocal microscope used was the KEYENCE VK-X150, which employs a 658 nm red semiconductor laser for imaging and measurement, providing a height range of 7 mm and a maximum resolution of 0.005 μm. The vk analysis software was used to process the collected data, providing three-dimensional morphology and roughness values for the measured area.

The workpiece in the HCAF is fully submerged in water, operating under a submerged water jet environment. Simulations were conducted using ansys fluent. Based on the structure of the mixed-flow chamber and flow channel workpiece, as well as the system's operating mechanism, a simplified two-dimensional geometric model was created, as shown in Fig. 6. The computational fluid domain parameters are listed in Table 2.

A mesh independence verification was performed using the maximum velocity in the flow field as a reference value. The mesh independence verification results are presented in Table 3. As the number of mesh elements increases, the maximum velocity in the flow field stabilizes. Using a global mesh size of 0.2 mm as the baseline, the relative error of the maximum velocity in the flow field remained within 0.5% for global mesh sizes of 0.3 mm and 0.15 mm. The accuracy of the numerical simulation no longer varied with mesh refinement. The final selected mesh size was 0.2 mm globally and 0.008 mm locally, resulting in a total of 733,921 mesh elements.

The SST k–ω turbulence model was selected for the numerical simulation. The mixture multiphase flow model was chosen, with slip velocity ignored. The Schnerr–Sauer cavitation model was applied, with bubble size set to 1 × 10⁻⁵ m and bubble nucleus density of 1 × 10¹³ nuclei/m³.

The ambient pressure for the numerical simulation was set to standard atmospheric pressure. Pressure inlet 1 was designated as the cavitation jet pressure (P_u1) and set to 15 MPa. Pressure inlet 2 (P_u2) was varied between −0.04 MPa and 0.06 MPa across different cases. The pressure outlet was set to 0 MPa. The computational domain walls were defined as no-slip stationary walls, with scalable wall functions applied for near-wall treatment.

The numerical simulation employed a pressure–velocity coupling approach, with the coupled solver selected. The pseudo transient was applied to enhance the convergence. First-order upwind schemes were used for the momentum, swirl velocity, volume fraction, turbulence dissipation, and turbulent kinetic energy equations. The residual convergence criterion was set to 1 × 10⁻⁵, with a maximum iteration count of 2000. The boundary condition parameters are listed in Table 4.

Figure 7 presents the surface roughness measurements of the top and bottom surfaces of the channel after polishing. Analysis of the surface roughness (R_a) curve indicates that the best polishing effect was achieved at a pressure of P_u2 = −0.04 MPa. In the 5–55 mm section of the flow channel, the surface roughness (R_a) was reduced from an average of 15 μm to approximately 1–2 μm, achieving an improvement of about 90%. In the 65–75 mm section, the roughness was reduced to approximately 7 μm, with an improvement of about 50%. In the 85–95 mm section, which is closer to the outlet, only a slight improvement in the surface roughness was observed. Furthermore, comparing the polishing effects under different abrasive pressures reveals that the HCAF process is highly sensitive to variations in abrasive pressure, with P_u2 = −0.04 MPa identified as the optimal process parameter.

Since the data at an abrasive flow pressure of P_u2 = −0.04 MPa are representative of this study, optical images and 3D profile images of the internal surface of the channel were extracted for the surface morphology analysis.

Figure 8 illustrates the longitudinal surface morphology of the bottom surface of the flow channel under an abrasive flow pressure of P_u2 = −0.04 MPa. As the distance increases, the polishing performance of the HCAF gradually decreases. In the 5–35 mm section of the flow channel, the original surface morphology was completely removed, revealing a smooth and uniform surface. From the 45 mm region onward, the original surface morphology begins to reappear. The 3D profile images reveal that the region near the inlet has a larger arc radius due to greater material removal, resulting in improved surface quality. Conversely, the region near the outlet experiences less material removal, retaining a significant portion of the original morphology.

Compared to the bottom surface of the flow channel, the top surface is more challenging to polish due to its lower initial surface quality, the top surface retains more of the original morphology after polishing due to its lower initial surface quality. Additionally, the reduction of surface profile peaks is more limited, as shown in Fig. 9.

The simulation results in Fig. 10 illustrate the pressure distribution inside the flow channel under different abrasive flow pressures. By limiting the maximum displayed pressure in the flow field to a relative pressure of 0 MPa, emphasis is placed on regions where the pressure is below 0 MPa. The minimum pressure in the flow field is −97,785 Pa (relative pressure), corresponding to an absolute pressure of 3540 Pa, which is the saturated vapor pressure of water under the current operating conditions. The dark blue regions in the figure indicate areas near the minimum pressure in the flow field, where cavitation bubbles can form and remain stable. Additionally, a pressure transition point exists in the flow channel under all abrasive flow pressure conditions. The pressure in the flow channel rises sharply from a lower pressure to atmospheric pressure, and as the abrasive flow pressure (P_u2) increases, this transition point gradually shifts closer to the flow channel inlet.

Figure 11 presents the simulated cavitation bubble volume fraction along the near-wall region of the flow channel under different pressures. Under all abrasive flow pressures (P_u2), the cavitation bubble volume fraction on the channel wall rapidly increases from approximately 0.1, reaching a peak of around 0.75 at approximately 3 mm. It then gradually decreases along the flow path before abruptly dropping to zero at different positions. In addition to reducing the cavitation bubble distribution area within the flow channel, an increase in the abrasive flow pressure also accelerates the cavitation bubble collapse rate. At an abrasive flow pressure of P_u2 = −0.04 MPa, the cavitation bubble volume fraction exhibits the slowest decline rate, which increases progressively with rising pressure.

Figure 12 presents the surface roughness measurements of the bottom and top surfaces of the flow channel after polishing. Regardless of whether surface roughness R_a or R_z is considered, the roughness curves generated using the Φ0.4 mm cavitation jet nozzle and the Φ1.0 mm nozzle do not intersect. Furthermore, the Φ1.0 mm curve is consistently lower than the Φ0.4 mm curve, indicating superior surface quality. A comprehensive analysis of the surface roughness measurements and morphology of the bottom surface of the flow channel indicates that the Φ1.0 mm cavitation jet nozzle exhibits significant advantages over the Φ0.4 mm nozzle in terms of the polishing performance. Under identical polishing conditions, the Φ1.0 mm nozzle achieved greater surface quality improvements and significantly enhanced polishing efficiency.

Similarly, by varying the nozzle diameter, the simulation results in Fig. 13 illustrate the pressure distribution inside the flow channel under different cavitation jet nozzle diameters. The pressure in the flow channel regions corresponding to both cavitation jet nozzle diameters is below the saturated vapor pressure, allowing cavitation bubbles to form and remain stable.

A comparison of the inner surface morphology of the top and bottom regions before and after HCAF, as shown in Fig. 14, reveals that the initial overhanging top surface exhibits significantly poorer quality than the bottom region, with numerous partially unmelted particles distributed across the surface. After optimization of the processing parameters, these unmelted particles were completely removed, and dense abrasive micro cuts were observed on the polished surface. However, porosity defects remained on the upper surface region, which are difficult to eliminate during the HCAF process. These results indicate that the HCAF is highly effective at removing protrusion-type defects, such as partially unmelted particles and slag, from the inner surfaces of additively manufactured channels. However, its removal efficiency is relatively low for fine cracks and similar defects.

Additionally, detailed surface morphology results under different processing parameters (abrasive pressure and nozzle diameter) are shown in Fig. 15.

Cavitation occurs when the static pressure of a fluid drops below its vapor pressure at a given temperature [19]. During the HCAF process, cavitation collapse near the machining surface generates localized high-speed microjets with velocities ranging from 200 m/s to 700 m/s. These high-speed jets repeatedly implode on the workpiece surface, eroding the material. Additionally, rigid abrasives near the cavitation implosion point are accelerated by the microjets, impacting the solid surface at higher velocities and enhancing erosion efficiency. Notably, in a static field, microjets and shock waves generated by the cavitation implosion can accelerate abrasives up to 40 m/s [20,21]. Therefore, cavitation intensity has a significant impact on the machining strength of the abrasive cavitation jet.

Figure 16 presents the results captured by the high-speed camera system. To better observe the visual behavior of cavitation bubbles and avoid interference from abrasive particles, a zero-abrasive concentration condition was used. As the abrasive flow pressure (P_u2) increases, the collapse rate of cavitation bubbles accelerates significantly, reducing their proportion in the flow channel. Compared to the flow channel being almost filled with cavitation bubbles at P_u2 = −0.04 MPa, at P_u2 = 0.06 MPa, cavitation bubbles are present only in a small region near the inlet. Based on the working mechanism of the HCAF, cavitation bubble collapse accelerates abrasives, enhancing the cutting force. Therefore, the distribution of cavitation bubbles directly influences the polishing results. The distribution of cavitation bubbles suggests that the polishing effectiveness of the HCAF decreases progressively with increasing abrasive flow pressure (P_u2) and declines along the flow direction.

In the equation, P_out represents the outlet pressure, which is set to standard atmospheric pressure in this study, P_v represents the liquid's saturated vapor pressure, and V_throat is the velocity at the throat, obtained as the average velocity across the throat cross section in the calculations.

The different initial cavitation numbers under different abrasive inlet pressures were obtained through calculation，as shown in Table 5.

Calculations reveal that when the water jet pressure remains constant, an increase in the abrasive flow pressure from −0.04 MPa to −0.06 MPa results in a slight increase in the cavitation number, with a growth rate of no more than 0.13%. Therefore, it can be concluded that the abrasive inlet pressure has a negligible effect on the incipient cavitation number.

Similarly, by varying the nozzle diameter, the calculated incipient cavitation number at the nozzle throat is presented in Table 6.

As the nozzle diameter increases, the cavitation number decreases significantly, with a reduction of 2.07%, following the trend described in the relevant literature. This indicates that variations in the nozzle size significantly affect the incipient cavitation number.

Additionally, during the experiment, the size of the flow channel significantly influences the polishing performance. A smaller flow channel diameter causes a rapid pressure rise under an adverse pressure gradient, leading to quicker cavitation bubble collapse, which prevents the bubbles from reaching a sufficient concentration for effective polishing. A larger flow channel diameter results in a rapid decrease in the flow velocity, thereby reducing the cutting force of abrasives and the synergistic polishing effect between the cavitation bubbles and abrasives, ultimately decreasing the polishing efficiency. The appropriate cavitation jet pressure for different flow channel diameters was analyzed through simulation, with the results presented in Fig. 17. The simulation results were compiled to extract the minimum cavitation jet pressure under the conditions where the near-wall flow velocity exceeded 30 m/s and the cavitation bubble volume fraction along the near-wall region remained above 0.4. A fitted curve was plotted in Fig. 18, revealing that variations in the cavitation jet pressure can, to some extent, be correlated with changes in the flow channel diameter.

Based on the previous polishing and simulation results, the optimal abrasive flow pressure parameter (P_u2) is −0.04 MPa. A comprehensive analysis of the polishing results indicates that the abrasive cavitation jet polishing process is highly sensitive to variations in abrasive flow pressure, with minor pressure fluctuations leading to a rapid decline in polishing efficiency. Combined with the previous analysis, the impact of abrasive flow inlet pressure on the incipient cavitation at the nozzle is negligible. Additionally, as the abrasive flow inlet pressure increases, the abrasive flow inlet flowrate also increases, exhibiting an approximately linear relationship. The relationship between the abrasive flow pressure and abrasive flowrate is presented in Fig. 19.

R₀ represents the maximum bubble radius at the onset of collapse, and ρ represents the liquid density [23]. Previously, it was observed that the abrasive inlet pressure has minimal impact on the overall flowrate. Therefore, assuming that the flow velocity remains constant across different abrasive flow pressures (−0.04 MPa to 0.06 MPa), the effect of abrasive flow pressure on the cavitation intensity is primarily reflected in the duration of high-pressure fluctuations (ΔT) at the branch junction. The pressure contour plots indicate that at 0.06 MPa, the duration of high-pressure fluctuations is significantly longer than at −0.04 MPa.

Additionally, an increase in abrasive flow pressure allows excessive fluid to enter the mixed-flow chamber, disrupting the flow field.

The effects of abrasive inlet pressure and nozzle size on the cavitation intensity in the mixed-flow chamber region were analyzed. Based on the simulation results, the underlying mechanisms were explained, and the optimal cavitation inlet pressure for different polishing channel diameters was determined. The above results were validated through observations of post-polishing surface roughness and high-speed camera imaging. The conclusions are as follows:

• The surface roughness results after polishing indicate that the best polishing performance is achieved at an abrasive inlet pressure of P_u2 = −0.04 MPa, while other pressure conditions result in less effective polishing. A nozzle throat diameter of 1 mm results in better polishing quality than 0.4 mm. Additionally, the closer the inner surface of the flow channel specimen is to the inlet, the better the polishing effect, achieving up to a 90% improvement in the surface quality.

• The simulation results indicate that flow channel diameter significantly affects the polishing performance. Both larger and smaller diameters require higher cavitation jet intensity, suggesting an optimal process window for the HCAF.

• Analysis of the simulation results reveals the impact of abrasive inlet pressure on the cavitation intensity within the mixed-flow chamber. The abrasive inlet pressure has only a minor effect on incipient cavitation at the nozzle throat, with an impact of approximately 0.13%. However, it significantly influences the distribution of high-pressure regions within the mixed-flow chamber. The high-pressure regions accelerate cavitation bubble contraction and collapse, reducing bubble lifespan and weakening the intensity and persistence of the cavitation effect.

• The National Key Research and Development Program of China (Grant No. 2022YFB4602502).

• The National Natural Science Foundation of China (Grant Nos. 52222503 and 52005437).

• Zhejiang Pioneer Plan Project (Grant No. 2025C01050).

There are no conflicts of interest.

The datasets generated and supporting the findings of this article are obtainable from the corresponding author upon reasonable request.