Disc and fiber gain ground
Laser welding efficiency and cost is compared among CO2, Nd:YAG, fiber, and disc lasers
Laser welding efficiency and cost is compared among CO2, Nd:YAG, fiber, and disc lasers
Stanley L. Ream
Healthy competition has led to tremendous improvement and innovation in laser design, and lasers are now an indispensable and cost-effective solution to many manufacturing challenges. The cost of laser power has continued to fall (in real dollars), and laser beam quality has continued to improve for all industrial lasers. Now, with the emergence of high-power disc and fiber lasers, the competitive landscape has expanded, and potential users are faced with new questions.
In a literal sense the term, “efficiency” usually applies to a dimensionless number, a fraction or percentage, that relates an actual value to a theoretical or optimal one. By this measure the phrase “laser welding efficiency” would be the ratio of the theoretical amount of energy required to melt that volume of metal compared to the delivered laser energy. To produce such a ratio the theoretical energy value must first be established. For this investigation the subject material is low carbon steel, specifically AISI-SAE Grade 1008.
The energy required to bring a volume of steel from room temperature to the melting temperature comprises “real heat,” transformation heat, and the heat of fusion (melting). Real heat is a function of change in temperature and heat capacity (specific heat, Cp), which varies considerably from room temperature to melting. The heat of transformation can also be captured as a change in Cp. Much of this information is available in materials handbooks,1,2 but rarely is all the informaiton available from room temperature through melting. A compilation/extrapolation of published Cp values for 1008 steel is shown in Figure 1. Using these Cp values the calculation of energy required to raise a gram of 1008 steel from room temperature up to the melting temperature yields a value of 1131 J/gm. [As a benchmark the frequently referenced AWS publication3 by Swift-Hook and Gick yields a value of 918 J/gm, using a single average value of 620 J/kg for Cp and no consideration for the heat of fusion.] Published values for the heat of fusion of mild steel are much more readily available and consistent, about 272 J/gm. In summary, the calculated “absolute” energy required to melt 1008 steel is approximately 1403 J/gm or 11 J/mm3 of steel.
FIGURE 1. Specific heat of 1008 steel as a function of temperature.
Unfortunately, this detailed information is rarely available for many engineering materials. Therefore, the absolute efficiency will not be highlighted in this analysis. Rather, the term “laser welding efficiency” will be used here to describe melt volume per unit of laser energy input or mm3/kJ, a term that does not depend on an absolute value.
Efficiency data generation and analysis
To determine laser welding efficiency (as defined above) three values must be obtained:
- Laser power on the work (kW)
- Travel speed (mm/s)
- Weld nugget area (mm2)
In previous work4 (1995) on this topic more than 100 sets of such data were collected and analyzed in a search for relationships between efficiency and power, travel speed, aspect ratio, etc. From Figure 2 it can be seen that there was considerable scatter in this data. Nevertheless, the analysis concluded that the maximum, measured, laser welding efficiency was approximately 50-60 mm3/kJ and this could be accomplished over a range of powers and travel speeds.
FIGURE 2. Laser welding efficiency versus power (Laser’95, Munich).
In the work performed at EWI it was possible to establish much tighter control over the welding and the data analysis. Three of the four lasers under consideration were available at EWI:
- PRC, 6kW CO2 laser
- IPG, 4kW fiber laser (beta unit at 3.8 kW)
- TRUMPF, 4kW, lamp-pumped Nd:YAG laser
Data from the fourth laser, a 4kW disc laser, was generated by TRUMPF in Germany. The basic set-up conditions for the four lasers are shown in Table 1. Off-axis gas shielding (helium for the CO2 and argon for the others) at 24 l/min (50 cfh) was used at EWI.
FIGURE 3. Method for weld nugget area analysis.
Metallographic weld cross sections were prepared and imaged for area analysis, which was accomplished with AutoCAD assistance (see Figure 3). Speed/penetration curves for the several lasers are provided in Figure 4.
FIGURE 4. Laser welding penetration in mild steel.
Laser welding efficiencies were calculated based on actual power measured at the work:
Efficiency = Nugget area × Travel speed/Power on work = (mm3/kJ)
The summary of laser welding efficiencies is presented graphically in Figure 5. Note that only two data points are available for the TRUMPF disc laser.
FIGURE 5. Laser melting efficiency in mild steel.
In contrast to the 1995 study the current work produced much more consistent data. All four lasers produced high efficiency in the mid-speed range, with values to 60 mm3/kJ, which equate to 66% absolute efficiency, based on the earlier Cp and melt energy discussion. This is notably higher than Swift-Hook and Gick’s prediction of 48% maximum efficiency, but about half of this difference in absolute efficiency can be attributed to differences in applied Cp and the absence of heat of fusion values in their analysis. In any event it is clear that all four of these lasers can produce nearly optimal melting efficiencies.
Welding cost calculations
This is very likely the most complex, misunderstood, and misrepresented area of industrial laser discussions. Too often the evaluations of laser welding cost only include a few variables, such as initial investment, running cost, and maintenance. Rarely are all the operational conditions taken into account. Since the full cost of production laser welding would be determined in large part by the overall system costs, only the laser, chiller, and beam delivery components costs and their maintenance will be analyzed here.
For this analysis there has been a “sanitization” of the cost data. In other words no exact laser prices were used, and long-term replacement parts intervals and costs were estimated in some cases. Laser manufacturers were consulted and assured that specific cost element information would not be presented here.
The basic operational view for this analysis is that of a high-production plant, operating three 8-hr shifts/day, 5 days/wk, 240 shifts/yr. Lunch, breaks, and maintenance time are deducted from shift availability, resulting in a nominal 80% system availability for each case. Maintenance, at $28/hr, includes laser equipment as well as beam delivery, focusing optics, and welding process issues, which account for much of the total maintenance time. Major maintenance intervals are conducted and calculated on an off-shift basis. The cost of the system operator is not included here. That cost would be carried by a “system” cost analysis, which is independent of the welding source. The depreciation and interest values used in the summary analysis are averaged over the eight years. Floor space cost is $20/ft2/yr, including non-laser utilities. Electric power cost is $0.06/kW-hr. Gas costs are typical.
Operating cost summary
The summary cost table is presented in Figure 6. There are many ways to look at these results, and the relative heights of the columns can be significantly altered by changes in electric cost and major replacement item costs and intervals. The most problematic question concerns anticipated diode life in disc and fiber lasers. There is simply insufficient data at present to know the eventual diode replacement costs and lifetimes, but the estimates embedded here are based on current experience and ongoing warranties.
FIGURE 6. Laser operating cost summary.
Welding cost summary
As part of the welding efficiency calculations the values for melting rate (mm3/s) were established for each bead-on-plate case. Using typical values for the mid-speed welding performance and the cost values from Figure 6, it is possible to express the average cost of welding, as shown in Table 3.
It must be re-emphasized that these are relative numbers based on a given set of assumptions and that they do not include any laser system components or labor. Just as importantly, they represent laser bead-on-plate welding, which is only reasonably close to partial penetration butt or lap welding of tightly fitted parts. Nevertheless, these values serve to illustrate just how relatively inexpensive the laser welding process can be.
The specific cost/cm3 value is just one measure of the value of different lasers in a given welding situation, and there are many other valuation factors that must be considered. And, as noted above, there are also estimated elements embedded in this costing examination. These elements will evolve and change with time, as will the absolute values in this table. Still, there are several broad conclusions that may be drawn.
- The welding efficiency of modern industrial lasers exceeds previously held beliefs concerning maximum theoretical limits.
- Initial laser cost may not dominate the overall cost analysis.
- Near-infrared lasers appear to offer slightly higher welding efficiency than CO2.
- Welding efficiency values can be used to assist in laser process analysis.
- For welding requiring only simple part motion, CO2 lasers present the most economical solution.
- In the not-to-distant future it appears that Nd:YAG lasers will likely yield to fiber and disc lasers.
Stanley L. Ream is Laser Technology Leader at Edison Welding Institute (EWI), Columbus, OH. Contact him at firstname.lastname@example.org or visit www.ewi.org.
- American Society for Metals Handbook, Vol. 1, 10th edition, p. 198.
- MatWeb, Materials Property Data, AISI 1010 Steel.
- Swift-Hook, D. T., and Gick, A.E.F., 1973, “Penetration Welding with Lasers,” Welding Journal 52(11):492s-499s.
- Ream, S. R., Bachholzky, E. J., 1995, “Analyzing Laser Butt Welding Efficiency,” 12th International Congress (Laser’95), pp. 247s-256s, Munich, Germany.
- Fuerschbach, P.W., 1996, “Measurement and Prediction of Energy Transfer Efficiency in Laser Beam Welding,” Welding Journal, Jan.’96, pp. 24s-34s.