Conical or Supersonic - Which Nozzle is better for Laser Cutting
J. Fieret and C. Rand
The laser cutting process has been extensively researched and modeled over the last couple of decades, especially the cutting assist gas interaction with the metal, both for active and passive assist species. The importance of shock discs was recognized early.1 This article addresses a deceptively simple question: are conventional (conical) nozzles or supersonic nozzles better for laser cutting? Given the large cost difference between the two, this seems a relevant question.
Nozzles
From a gas dynamics point of view, these nozzles are very different. The supersonic minimum length nozzle is probably the most-researched type, providing the best compromise of shock-free flow, nozzle length, and ease of manufacture. This nozzle has a conical contracting section, followed by an expansion part that is matched to the expansion of the compressed gas at a particular pressure, resulting in smooth flow without significant shocks. True Laval nozzles, which have a shaped contracting section and a smooth throat section, can produce a completely shock-free flow, but they are impractical because of their high cost as no mass manufacturing technique exists.
Sonic nozzles (see Figure 1) are the commercial standard for industrial laser cutting. When operated at the critical pressure ratio of p0/pB = 1.89 the resultant flow has a Mach number of 1.0 (p0 is the pressure applied to the nozzle, and pB is ambient pressure). When using bar as unit of pressure, this ratio is numerically approximately the same as the pressure applied to the nozzle, because ambient pressure is almost exactly 1 bar.
Laser cutting with such low assist gas pressures is rarely performed on thin section material, as gains in processing speed and quality are obtained by operating at higher inlet pressures. Increasing the inlet gas pressure causes the free jet to become under-expanded at the nozzle exit resulting in shock wave formation (the term under-expanded means that on exiting the nozzle, the gas flow wants to expand further). This phenomenon is indicated in Figure 2, which shows the calculated and real pressure pattern in the gas flow downstream from the nozzle aperture, for an applied nozzle pressure of 7.8 bar.
The differences between the two nozzle types are apparent: the supersonic minimum length nozzle shows almost no shocks (the structures that are visible are comparatively weak expansion and compression waves), indicating the (near) absence of pressure variations, while the other nozzle shows strong shocks and hence large pressure variations. Supersonic minimum length nozzles are designed using the method of characteristics (a computational fluid dynamics algorithm).
While the differences in flow characteristics are obvious, the question remains whether the smooth supersonic flow from the minimum length nozzle has advantages to the cutting process.
Melt removal mechanisms
The capability to eject the molten metal from the cut kerf is dependent on the drag from the gas assist jet along the cutting front. There are two types of drag: that due to tangential stresses (viscous drag), and that due to normal stresses (pressure drag). Viscous drag is most prominent where the surface area parallel to the flow direction is large compared to the projected area normal to the flow. Pressure drag arises when the cut front is at an angle to the gas jet where a shear force will act upon it and a pressure gradient will be formed. Viscous drag and pressure drag are of the same order of magnitude in laser cutting, and are the main forces of melt and vapor removal from the kerf.3
Kerf flow conditions
According to the theory of gas dynamics, the best condition for material removal from the cut is when the exit pressure at the bottom of the kerf is equal to the ambient pressure. If it is much higher than the ambient pressure, not only is the viscous force reduced, due to the decrease in gas velocity approaching the kerf bottom, but also the flow leaving the kerf will rapidly accelerate due to supersonic expansion. As a result, the molten material is ejected in divergent directions not along the tangential direction of the cut front. Some molten metal is forced towards the sidewalls and forms dross clinging to the bottom of the cut edges. If the exit pressure is lower than the ambient pressure (this is possible when the flow in the kerf is supersonic), strong shock waves will be formed at the exit. These shock waves not only result in energy loss but also a sharp increase in the exit pressure and a sudden reduction in gas velocity and hence poor cutting. Therefore, an exit flow condition with maximum gas velocity and a pressure equal to the ambient pressure is desired to achieve a cut edge without dross.4
Industrial use of nozzles for laser cutting is generally limited to sonic (conical) nozzles operating at high inlet gas pressures and hence with under-expanded jets. To look at the effects of increasing the inlet pressure ratios, a sonic nozzle and a supersonic minimum length nozzle were modeled at nozzle pressures of 1.89, 3.7, 7.8, and 17 bar.
Sonic nozzles cutting performance
Figure 3 shows computed Mach number contour plots in the cut, sideways viewed (blue means low Mach number). The gas jet points vertically down, and moves from right to left (cut direction). In the first instance, consider the fully expanded jet (nozzle pressure of 1.89 bar) exiting from a sonic nozzle impinging upon the workpiece. As the flow hits the surface and enters the kerf, a normal compression wave develops at the workpiece surface. In the kerf below the compression the flow detaches from the kerf front and accelerates again as it expands into the kerf (yellow region). Following this expansion there is a flow compression, as the flow tries to reach a steady state, through a compression wave. This is not a good condition for optimum melt removal, as the pressure force, and hence the drag, is very low.
As the pressure is increased, the compression wave on the workpiece surface becomes a shock and the flow separation from the kerf front reduces as the flow speed inside the kerf increases. At a pressure of 17 bar, the flow is not only supersonic throughout the kerf, but speeds up even as it leaves the bottom of the kerf (under-expanded). As a result, the molten material would be ejected in divergent directions not along the tangential direction of the cut front. Some molten material would be forced towards the sidewalls to form dross clinging to the bottom of the cut edges.
Calculations and experiments show5 that at a nozzle pressure of approximately 10 bar the flow is fully expanded, and hence at this pressure the cutting process is optimized, with no adherent dross. The numerical results were verified with cutting tests using a 2.25kW CO2 laser and a 2.5in focal length lens. Having an M2 value of 1.2 this gave a spot size of about 75µm and a power density of 4x107 W/cm2. When cutting 1mm mild steel with a nitrogen assist gas the optimum pressure was about 10 bar. This not only produced the fastest cuts, but also yielded the smoothest surface finish with the least amount of dross, in agreement with the model.
There was however one discrepancy: according to the flow conditions, higher nozzle pressures should result in higher cutting speeds whereas in fact the experiments indicate that this is not the case. The cause of this is the formation of a plasma plume above the workpiece at higher pressures, which absorbs the laser beam.
Supersonic nozzle cutting performance
The flow conditions in the kerf for the supersonic nozzle are remarkably similar to that of the sonic nozzle, especially at low pressure, even though the nozzle flow conditions are totally different. As shown in Figure 3, at 1.89 bar the conditions in the kerf are identical, as can be expected because the impact velocity in both cases is sonic. At higher pressures the shock over the workpiece is stronger than for a conical nozzle, because the impact flow velocity is higher (fully expanded supersonic flow). Shock waves destroy kinetic energy (turning it into heat), resulting in a slower flow inside the kerf than for conical nozzles, which is quite contrary to intuition. The supersonic nozzles also generate less expansion of the flow within the kerf, which reduces the force upon the kerf front, which acts upon the molten layer. It can therefore be expected that supersonic nozzles will in fact result in a lower maximum cutting speed, at lower nozzle pressure. The model indicates a fully expanded flow at the bottom of the kerf at around 8 bar, at which dross-free cutting can be expected.
In the laser cutting results considered here, very high power intensities were generated, leading to intense plasma formation above the workpiece. Plasma formation above the workpiece absorbs some of the laser radiation reducing the energy available to generate the melt. The flow from the supersonic nozzles has been expanded before it exits the nozzle resulting in lower pressure and density above the workpiece than the sonic nozzle. This reduction in density limits the plasma formation allowing more of the laser power to reach the workpiece. Such high laser power intensities are not usually seen in laser cutting; however as laser powers increase the issue of plasma shielding will become more of an issue.
Cutting trials2, 5 with the laser setup confirm these expectations, as indicated in Figure 4.
Conclusions
The main conclusion is that, for thin steel sheet cutting, supersonic nozzles do not improve the melt removal conditions, as the advantages of the supersonic flow are completely lost on impact with the workpiece surface. There may however be an advantage with thick section cutting, where the kerf can be much wider.
Additionally, cutting speeds are not limited by the melt removal rate. Increasing gas pressure does not necessarily increase cutting speeds, as at high pressures the formation of plasma above the kerf reduces the incident laser energy on the workpiece limiting melt production and cutting speeds.
For optimal cutting conditions there is a need to deliver a gas jet with the ability to fully expand to atmospheric pressures at the kerf exit. Dross-free cuts are achieved when the flow out of the kerf is fully expanded. For thin sheet, both conditions are most easily achieved with a conventional conical nozzle.
References
- Fieret, J., Terry, M.J., Ward, B.A., “Aerodynamic interactions during laser cutting,” Proc. SPIE Int. Symposium on Optical and Opto electronic Applied Sciences and Engineering, Conference on Laser Processing: Fundamentals, Applications and System Engineering, Quebec, Canada, 2-6 June 1986.
- Rand, C., “The Study of Sonic and Supersonic Jet-Kerf Dynamics in Optimised Laser Cutting”, PhD Thesis, The University of Liverpool, 2004.
- Vicanek, M. and Simon, G., “Momentum and heat transfer of an inert gas jet to the melt in laser cutting,” J. Phys. D: Appl. Phys. 20, pp 1191-1196, 1987.
- Duan, J., Man, H.C., and Yue, T. M., “Modelling the laser fusion cutting process: III. Effects of various process parameters on cut kerf quality,” J. Phys. D: Appl. Phys 34, pp 2143-2150, 2001.
- Rand, C. and Fieret, J., “Advances in modern laser cutting,” ALAC 2006, Novi, MI (US), 18-21 Sep 2006.
The authors are with The Linde Group Guildford, Surrey, UK. Jim Fieret can be contacted by e-mail [email protected].