Beruflich Dokumente
Kultur Dokumente
WELDING RESEARCH
FCAW, and three GMAW methods. The processes studied were compared for costs, range of common welding processes.
including relative labor costs. Results indicate the Cr6+ in the fume varied widely, from
a low of 2800 to a high of 34,000 ppm. Generation rates of Cr6+ ranged from 69 to Common Welding Configurations
7800 μg/min, and Cr6+ generation rates per unit of wire ranged from 1 to 270 μg/g.
The results of field study were similar to the findings in the laboratory. The Cr6+ (ppm) Analysis of welding-based hazards is
in the fume did not necessarily correlate with the Cr6+ generation rate. Physical prop- dependent on an understanding of the
erties were similar for the processes, with mass median aerodynamic diameters rang- range of welding processes and condi-
ing from 250 to 336 nm, while the FCAW and SMAW fumes were larger (360 and 670 tions. More than 80 different welding
nm, respectively). Conclusion: The pulsed axial spray method was the best choice of processes (Ref. 6) are commonly found,
the processes studied based on minimal fume generation, minimal Cr6+ generation, but most welding is done with electrical
and cost per weld. This method is usable in any position, has a high metal deposition arc welding processes. The most preva-
rate, and is relatively simple to learn and use. lently used variations, based on materials
usage, are shielded metal arc welding
(SMAW) ~ 45%; gas metal arc welding
(GMAW), ~34%; and flux cored arc
Introduction steel may account for up to 5% of welding
welding (FCAW), ~17% (Ref. 5).
(Ref. 3), and is an important segment of
The SMAW process has the simplest
Welding is a major occupational activ- some industries, such as food processing
equipment requirements: a power supply,
ity in the United States and worldwide, equipment manufacturers, chemicals
an electrode holder, welding rods, and a
and includes workers in manufacturing, manufacturing, and shipbuilding.
ground clamp. The welding rods have a
construction, and a number of other in- However, welding on stainless steel
coating over the filler metal rod that pro-
dustrial sectors. In excess of 462,000 U.S. presents an additional occupational-risk
vides a shielding environment to minimize
workers do some welding as part of their exposure; the known carcinogens hexva-
degradation of the weld by atmospheric
duties (Ref. 1), and about two-thirds of lent chromium (Cr6+) and nickel (Ni) have
oxygen or nitrogen. The GMAW process
these workers are in manufacturing indus- been identified in stainless steel welding
uses more complex equipment; besides a
tries. Welding produces a number of haz- fumes (Ref. 4). A review by the Interna-
power supply, it uses a gas-shielded gun
ards during operation, including fumes, tional Agency for Research on Cancer
and the electrode is a consumable wire of
gases, and physical agents such as heat and the desired filler metal fed by a motorized
ultraviolet and infrared radiation. Occu- feeder. The shielding gas is externally sup-
pational health studies indicate a number KEYWORDS plied from cylinders. Shielding gases
of occupationally related adverse health range from the completely inert argon
effects, such as lung disease (Ref. 2). Most Hexavalent Chromium (Ar), helium (He), and their mixtures to
welding operations are performed on low- Stainless Steel so-called active gases, which include car-
alloy or high-carbon steels, but stainless Gas Metal Arc Welding bon dioxide (CO2), Ar mixtures with CO2
(GMAW) or oxygen (O2), and other gas mixtures.
M. KEANE (mjk3@cdc.gov), S. STONE, B. Flux Cored Arc Welding These gases may have chemical interac-
CHEN, A. CUMPSTON, and J. ANTONINI are (FCAW) tions with the weld or fume. The FCAW
with National Institute for Occupational Safety Shielded Metal Arc Welding process uses equipment similar to
and Health (NIOSH), Health Effects Laboratory (SMAW) GMAW, but the wire electrode has an in-
Div., Morgantown, W.Va. J. SLAVEN is with Welding Fume ternal flux material for weld shielding; the
NIOSH and Indiana University, Indianapolis,
Ind. A. SIERT is with Xcel Energy, Denver, Colo.
process may be used with or without an ex-
ternal shielding gas.
Fig. 1 — American Welding Society-type welding chamber in use. Fig. 2 — Welding fume-generation rates for eight welding processes, in mg/min.
WELDING RESEARCH
Fig. 3 — Hexavalent chromium content in welding fumes from eight Fig. 4 — Hexavalent chromium generation rates for eight welding processes, in
processes, in mg/kg. μg Cr6+/min.
Ni, and the remainder Fe. Flux cored arc trol arc time.
welding used AWS A5.22 Class E308LT0-1 Operating variables are
wire, while SMAW generation used AWS shown in Table 1.
A5.4 Class E308L-16 3⁄16-in. (4.8-mm) rods.
Shielding gas was taken from pressure- Sampling Strategies
regulated cylinders at a flow rate of 19
WELDING RESEARCH
L/min. The welding material in the base- Fumes from the weld area
plates was ½-in.- (12.7-mm-) thick 304 were sampled through a 102-
stainless steel, 22-in.- (56-cm-) diameter mm filter at the top of the
disks. Operation with all shielding gases chamber at 200 L/min. The
and welding process types used similar filter material was
conditions, but was adapted to good weld- Hollingsworth and Vose Fig. 66+— Hexavalent chromium exposures inside welder’s helmets in
(East Walpole, Mass.) elec- μg Cr /m for five welding processes.
3
ing practice, generally as recommended by
welding machine manufacturers. All weld- trostatic medium (PE
ing operations were adjusted for good 13060NA), cut to fit the filter
bead appearance with good penetration of housing. The flow was measured with a hous-
the base plates and good toe wetting and mass flow meter (TSI, Shoreview, Minn.) ing had a short piece of 6-mm ID silicone
without undercut. Baseplates were ro- before sampling. After sampling was com- tubing, cut at 45 deg on the inlet end, and
tated at the desired rotary speed to pro- pleted, filters were removed from the a house vacuum was connected to the out-
vide a travel rate compatible with good housing, folded inward, weighed to the let end. Using a gentle blotting action,
welds. A commercial welding turntable nearest 0.1 mg, and put in sealed antista- most of the particulate fume was removed
was modified for external control by the tic polyethylene bags. from the filter. Sufficient quantity was col-
Labview (National Instruments, Austin, lected for analysis of Cr6+ and other met-
Tex.) program; rotary speed was detected Sample Recovery and Processing als, but quantitative recovery was not nec-
by an encoder on the output shaft and essary. After completion, the
input to the program, and the travel rate Welding fume particulate matter was polycarbonate filter was removed from the
displayed continuously during operation. recovered from the filters by gentle suc- housing over a tared 75 × 75-mm weigh-
Welding machine operation was initiated tion onto a 47-mm, 0.8-μm polycarbonate ing boat, and material brushed from the
under program control to precisely con- filter. The 47-mm stainless steel filter filter and housing interior with a #3
artist’s brush. The fume was treated with
an antistatic device at this point to prevent
losses. The dust was then ground in a
Table 1 — Operating Variables for Generated Welding Fume Samples metal-free apparatus to homogenize the
sample for replicate analyses. Preliminary
results indicated large differences be-
Shielding Gas Gas Flow Weld Mode Wire Feed Voltage (V) Current (A) tween replicate samples weighed from
cm/min material recovered from filters and subse-
Ar/CO2 99%/1% 19 L/min Short circuit 320 17–18 120 quently analyzed, with coefficients of vari-
Ar/O2 99%/1% 19 L/min Short circuit 320 17–18 125 ation typically >20% for replicate sam-
He/Ar/CO2 ples. The process below resulted in much
90%/7.5%/2.5% 19 L/min Short circuit 320 19.5 100 better precision for replicate samples,
Ar/CO2 99%/1% 19 L/min Axial spray 760 25 200 with coefficients of variation of typically
Ar/O2 99%/1% 19 L/min Axial spray 760 23 200 5–10%. The system used was a disposable
Ar/O2 98%/2% 19 L/min Pulsed spray 760 0.9 V trim (pulsed) 13-mm-diameter × 25-mm-long polyeth-
Ar/CO2 75%/25% 19 L/min FCAW 760 26 160 ylene vial with two 3.2-mm (1⁄8-in.) silicon
none none SMAW 361(a) ~20 150 nitride coated ceramic balls, shaken for 30
s in a Wig-L-Bug grinder. After grinding,
(a) Equivalent wire feed speed based on an equal mass consumption rate.
the material was antistatic treated again,
Field Study
mets, and the welders wore half-face respi- classify nano-size particles, the MOUDI are displayed in Fig. 3. The results are the
rators with P100 filters during welding. has rotating stages to obtain a nearly uni- arithmetic means ± standard error of
Short-term trials were conducted to main- form particle deposit on the collection three replicate samples for each of the
tain 8-h time-weighted average exposure substrates, which reduces particle bounce four welding runs. Results range from
levels below the OSHA Cr6+ permissible ex- and improves subsequent analysis. The 2800 ppm (GMAW-S using He/Ar/CO2
posure level (PEL). total flow rate for the impactors was 30 shielding gas) to 34000 (SMAW). The
L/min, and they were operated for 4 min Cr6+ generation rates were calculated as
Analysis for Cr6+ total sampling time. the product of the fume-generation rate
and the Cr6+ fraction in the fume, and are
Samples were treated and analyzed Statistical Approaches displayed in Fig. 4. Additionally, some of
using NIOSH Physical and Chemical the differences in total fume-generation
Analysis Method 7605, Hexavalent The data did not follow a normal dis- rate and Cr6+ generation rates were influ-
Chromium by Ion Chromatography (Ref. tribution, and were extremely skewed. enced by the different wire-feed rates nec-
9). The estimated limit of detection is 0.02 Therefore, the Kruskal-Wallis nonpara- essary to maintain optimal welds. The nor-
μg, and the method range is 0.05 to 20 μg metric test was used to look for statistically malized generation rate is related to a
of Cr6+. Five mL of extraction solution significant differences between groups. welder’s exposure for any given weld,
(3% Na2CO3/2% NaOH) were added to Statistical results were considered signifi- since the weld is not complete until suffi-
each 5-mg sample, and the tubes sonicated cant at a p-value of 0.05. cient metal is deposited to meet the re-
in a bath for 30 min. This procedure ex- quirements of that weld. Therefore, the
tracts both soluble and insoluble Cr6+ Results generation rates were renormalized with
present in the fumes. Samples were re- respect to wire feed rates; results are
moved and centrifuged for 15 min at 3500 Fume-generation rates for four runs shown in Fig. 5. The results normalized for
× g. The supernatant was transferred to are displayed graphically in Fig. 2. The re- wire-feed rates were calculated as the
25-mL volumetric flask and diluted with sults in Figs. 2–5 are presented as arith- product of the Cr6+ generation rate and
H2O. Samples were analyzed by ion chro- metic means ± standard error of the the reciprocal of the wire-feed rate in
matography using a Dionex HPIC-AS7
column with 250 mM (NH4)2SO4/100 mM
NH4OH mobile phase and a postcolumn Table 2 — Mass Median Aerodynamic Diameters (MMADs) and Geometric Standard
reagent (2.0 mM diphenylcarbazide/10% Deviations (GSDs) of Welding Fumes for Different Shielding Gases and Processes, from MOUDI
methanol/1N H2SO4) with absorbance de- Data
tection at 540 nm. Four concentrations of
standards were made from a certified Cr6+ Shielding Gas Process MMAD (nm) GSD
solution, covering a range of 0.4–4 μg/mL.
Ar/CO2 99%/1% Short circuit 340 1.46
Particle Size Distributions Ar/O2 99%/1% Short circuit 280 1.53
He/Ar/CO2 Short circuit 330 1.47
Particle size distributions of different 90%/7.5%/2.5%
welding fumes were determined by using Ar/CO2 99%/1% Axial spray 260 1.23
two Micro-Orifice Uniform Deposit Im- Ar/O2 99%/1% Axial spray 250 1.37
pactors (MOUDI and Nano-MOUDI, Ar/CO2 98%/2% Pulsed spray 320 1.58
MSP Models 110 and 115; MSP Corp., Ar/CO2 75%/25% FCAW 363 1.35
Shoreview, Minn.). By combining the two — SMAW 600 1.34
impactors, the particles were size-classi-
WELDING RESEARCH
in Fig. 6. The results are shown as box eral, but the very wide range of exposure torch cable can reach, unless obstacles,
plots of the observed exposure concentra- concentrations found suggests that fac- etc., are present. This may result in more
tions in μg/m3, with the median identified tors other than generation of Cr6+ at the fatigue, especially for vertical and over-
and the box spanning the 25th to 75th per- source are critically important in deter- head positions. Some training and testing
centile, and the whiskers encompassing mining exposures. While all of the may be necessary, but most GMAW
the entire range of observations. welders in the small group welded in the processes, including GMAW-P, are not es-
A typical particle size distribution is same shop on the same materials, the re- pecially difficult to learn or use.
shown in Fig. 7, including the curve for the sults indicate that a single observation for Some suggestions for easing adoption
best-fit unimodal model. All of the each process was dramatically higher would be to replicate the timing intervals of
processes also included substantial mass than the remaining exposures, for both current practice and keep welding rates
in the large, nonrespirable fractions >10 GMAW-spray and GMAW-P modes. (time for completion of a typical weld) sim-
μm, most likely associated with micro- There was no obvious explanation for ilar at the outset, even though GMAW
spatter from the welding processes. Sum- these anomalies, but some factor of work processes are often easier and faster than
mary data on geometric mean particle practice is probably responsible. SMAW.
sizes and respective geometric standard There have been a number of success-
deviations are shown in Table 2. ful approaches to reducing Cr6+ in weld- Recommendations
ing, including altered fluxes for SMAW
Discussion and FCAW (Refs. 12, 13), secondary For fume-generation rate, Cr6+ gener-
shielding gases for GMAW (Ref. 14), non- ation rate, and Cr6+ generation rate per
The results showed a wide range of Cr-containing welding alloy wires and unit of wire in this study, pulsed spray
fume-generation rates (a 14:1 ratio over- rods (Ref. 15), and chemical fume reac- GMAW was clearly the best of the
all, and 4:1 for GMAW), demonstrating tants (Ref. 16). This study concentrated on processes evaluated. A previous study in
there are clearly some best and worst reducing generation of Cr6+ at the process multiple industrial plants (farm machinery
choices for minimizing total fume expo- level, and could be used along with some manufacturing) confirms the clearly lower
sures. The results were somewhat greater of those other methods, such as Cr-free fume exposures in those facilities (Ref. 17)
than those found in similar studies (Refs. welding wire, fume reactants, and a sec- when using pulsed spray welding. The
8, 10, 11). A probable explanation is that ondary shielding gas. method has multiple practical advantages
the chamber used in this study was highly The various welding processes have in addition to the minimal Cr6+ and total
efficient with low losses and a minimal res- multiple associated cost factors that are an fume-generation rates: It is usable in any
idence time. Evidence for this includes the important issue in selecting a welding position, has low heat input that mini-
observation of very little loss of fume to process. Costs that need to be considered mizes warping, the high metal deposition
the interior of the chamber (< 1% of the include consumables (welding wire, rods, rate lowers labor costs, it is relatively sim-
collected mass for a given run), and the shielding gases, etc.), equipment costs, ple to learn and use, it has a noncritical
presence of substantial masses of large and labor costs per completed weld. working distance and good visibility, and is
(>10 μm and > 18 μm) fractions in the Table 3 presents relative costs in a non- only modestly more expensive than GMA
MOUDI sampler; these are easily lost by quantitative way, based on consumables welding machines with similar capacities.
sedimentation in large systems with long costs from typical industrial suppliers, and While there will be situations where
residence times and extended sampling manufacturers’ suggested prices on weld- pulsed spray mode may not be suitable, it
lines. ing equipment that have similar current would be a good choice in many applica-
The Cr6+ fractions (mg/kg or ppm) in capacities, but different process capabili- tions, especially where Cr6+ exposures
the fumes were comparable to those in ties. Comparisons in equipment costs are may be difficult to reduce by local exhaust
previous studies (Refs. 8, 10, 11): at best imperfect, and some process capa- ventilation or similar measures.
Disclaimer