**5. Conclusions**

included in this figure. It can be seen that the theoretical profile presenting the best matching

**Figure 6.** Comparison between calculated plasma velocity values at the axis for *c* = 0, *c* = 0.08 and *c* = 0.20, and the measured values of the axial velocity corresponding to light emitted from the arc central core. Taken from Mancinelli

**Figure 7.** Averaged theoretical axial plasma velocity over the emitting arc cross section, defined from the arc core (ax‐ is) to the ≈ 4000 K temperature line for the same *c* values presented in Fig. 3. The experimental values corresponding to an averaged emission over the whole emitting section of the arc are also shown. Taken from Mancinelli et al., 2011.

with the experimental data is that corresponding to *c* = 0.08.

78 Computational and Numerical Simulations

et al., 2011.

The modelling of dc arc plasma torches is quite challenging because the plasma flow is highly nonlinear, presents strong quantity gradients and is characterized by a wide range of time and length scales.

In the last years numerical plasma modelling has reached a state advanced enough to be of practical use in the study of cutting-arc processes. However, a self-consistent descrip‐ tion of the plasma starting from only macroscopic parameters (such as the geometry, current intensity, nature of the gas and type of employed materials, mass flow rate and/or some boundary conditions) has not yet been possible because of the lack of precise knowledge of some phenomena (electrode phenomena, radiation, turbulence, wall ablation, etc), which impose simplifications on the models. In particular, the practical use of cutting torch codes requires the introduction of some numerical coefficient whose value has to be obtained from a comparison between the model predictions and the experiment. For these reasons, the experimental validation of such models is of primary importance.

Assuming LTE conditions, the properties of the plasma that must be computed are the temperature, pressure and velocity fields. Numerical models for the plasma generated in cutting torches published during the last ten years have been validated using tempera‐ ture data derived from spectroscopic measurements in the nozzle-anode gap. It has been shown in this work that the plasma temperature is not the most appropriate quantity to validate numerical codes since it is not quite sensitive to changes in the model numerical parameters. Instead, it has been shown that the plasma velocity appears to be a more adequate quantity to perform such validation.

In order to realize this validation to such a sensitive variable as the plasma velocity, a 2- D model similar to those proposed in the literature was developed and applied to the same 30 A high-energy density cutting torch that was used in the velocity measurements recently published by some of the authors. Within the experimental uncertainties, it was found that a Prandtl mixing length turbulent parameter *c* = 0.08 allows to reproduce both the experimental data of velocity and temperature. However, this value has to be taken with caution, since that *c* value depends on the actual torch geometry, gas type and arc current. It can also be concluded that the simple Prandtl mixing length model is appropriated to predict the plasma characteristics in low-current cutting torches.
