Theoretical and experimental contributions on the use of plasma Arc installations in cutting different types of materials--Part 2.

Fagarasan, Cristian Codrut ; Popa, Marcel Sabin

1. INTRODUCTION

1.1 Factors affecting consumable life

In a properly functioning system, three key factors affect consumable life (The Hypertherm, sept. 2000): number of pierces, cut duration and material thickness.

These technologies by themselves, however, do not guarantee optimal consumable life. Other system problems still can cause unnecessary consumable wear. In most instances, these operating problems are fairly easy to identify and correct: premature or improper consumable changeout, pierce height, ramp-down errors in oxygen cutting systems, gas supply, coolant flow, and work cable connection.

2. CONTRIBUTIONS ON NUMERICAL OPTIMIZATION OF THE PLASMA ARC CUTTING PROCESS

2.1 Numerical optimization. Desirability values.

When studying material Stainless Steel, steel grade 5NiCr180, thickness 3 mm, it is designed to achieve optimization with the condition of minimum roughness (Fig. 1) and minimum cutting time (Fig. 2), regarding the degree of optimization, the points of prediction are 0,731 and 0,976.

Following the value curves indicated in Fig. 3, it is easy to see that they are more inclined towards the speed axis than the voltage axis, therefore suggesting an increased influence of the cutting speed parameter.

Once the steps for optimization are went through, the program offers a solution to obtain the minimum roughness value using the data provided. To obtain a minimum roughness, Ra = 19,47 [micro]m, and a cutting time value of t=15,21 s, the voltage must have the value of 170 V, the intensity 80 A and cutting speed of 975 mm/min.

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

It is noted that in Fig. 1, 2 and 3 Desirability term is represented, whose value varies on a scale from 0 to 1 (1 being the maximum). For material 5NiCr180 thickness 3 mm, that point's value is 0.976.

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

3. MATHEMATICAL MODELS RENDERING

A model rendering is a process of confirming that the model can approximate the outcome experimental values with a satisfactory level of accuracy. The main question is how to estimate the confidence level of a model. Mainly there are three techniques for validation of a model (Lazarescu et al., 2008): by graphical comparison, confidence interval method, and R 2 method (the calculation of determination coefficient).

Graphic comparison method is based on visual comparison by the same graphical representation of experimental and calculated results.

Confidence interval method involves analysis of the model parameter uncertainty in the process of validating the model and determines the level of confidence of the calculated response.

4. FRAMING WITHIN THE CONFIDENCE INTERVAL

The calculated F statistics are compared with the reference value F[alpha] (v1, v2), where v1 represents the number of degrees of freedom in the numerator (of the considered factor) and v2 the number of degrees of freedom of the error (in the denominator).

Analyzing the F statistics related to each mathematical model, (see ANOVA picture for response surface model for surface roughness and model of cutting time) in which the F (Fischer) statistic value is found, and considering the "Six Sigma philosophy" (Lazarescu et al., 2008), it was determined that the probability that the results obtained with mathematical models to be within the confidence interval [-3[sigma], 3[sigma]], is 99,95%.

5. CONCLUSIONS

It can be concluded that the material having a cut surface of the highest quality is undoubtedly steel grade S235JR (OL37-2k).

In terms of roughness, acceptable values (e.g. Ra = 5/15 [micro]m) were found for stainless steel, studied alloy was grade 5NiCr180. The third group of examined material in the study was aluminum Al99.5. It describes, following the cutting process and performed measurements, the higher values of roughness which is interpreted as a poor cut surface quality.

After presenting the methodology for establishing mathematical models describing the relationship between quality characteristics (cut surface roughness and cutting time) and process variables (cutting speed, voltage and generator current intensity), the mathematical models were tested using the ANOVA method, and by graphical comparison it was demonstrated that the solutions fits sufficiently precise the experimental results.

6. REFERENCES

Fagarasan Cristian-Codrut--Beitrage zur Messung von Plasmaparametern in eine Sputteranlage zur Beschichtung von Kleinteilen--Universitaet Stuttgart, IFF, 2003.

Fagarasan Cristian-Codrut--Studies and research on the use of Plasma Arc Installation in cutting different types of materials, Teza de Doctorat, Universitatea Tehnica Cluj-Napoca, 2009.

Lazarescu Lucian, Achimas Gheorghe, Ceclan Adrian Vasile, FEM-Simulation and response surface methodology for the analysis and prediction of cross section distornions in tube bending processes, 14th "Building Services, Mechanical and Building Industry days", International Conference, 30-31 Oct. 2008, Debrecen, Hungary,ISBN:978-963-473-124-5

Marinescu, N., Gavrilas, I., Visan, A., Marinescu, R.D., Prelucrdri neconvenfionale in construcfia de masini, Vol. II, Bucuresti, Editura Tehnica, 1993.

Popa, M.S.--Masini, tehnologii neconvenfionale si de mecanica find, editie bilingva, Editura U.T. Pres, Cluj-Napoca 2003.

Rossnagel, M. Stephen, Cuomo J., Jerome, Westwood D., William--Handbook of Plasma Processing Technology, Ed. Noyes Publications, New Jersey, USA, 1990.

The Hypertherm Inc., editia Sept 2000 a publicatiei The FABRICATOR[R], pp. 28-31, U.S.A., 2000.

Westkamper, E.; Warnecke, H.-J.; Gottwald, B. (Mitarb.)--Einfuhrung in die Fertigungstechnik. 4. uberarbeitete Auflage, Stuttgart; Leipzig: B