The Effectiveness of Controlled Spray Training

(near- and long-term) in the Fiber Reinforced Plastics Industry


S. J. Hall and J. R. Noonan

Indiana Clean Manufacturing Technology and Safe Materials Institute

Purdue University

West Lafayette, Indiana  47906


         The fiber reinforced plastics (FRP) industry accounts for a significant portion of Indiana’s economic activity.  In June of 1998, in response to newly proposed air pollution requirements and their potential effect on the FRP industry, the Indiana Clean Manufacturing Technology and Safe Materials Institute (CMTI) organized an industry consortium of FRP manufacturers and their suppliers to address the new requirements.  On December 8, 1998, The Greater Elkhart County Chamber of Commerce received a Strategic Development Fund grant from the Indiana Department of Commerce to assist the fiber reinforced plastics industry.  Indiana’s strategic development fund grants are awarded to companies and industry consortia to support training and technical assistance programs that make Indiana industries more profitable and competitive. The objective of the FRP grant is to implement an open mold, spray-operator training program, modeled after the Composites Fabricators Association (CFA) Controlled Spray Training program, and to evaluate the effect of the program.  The grant proposal was written by the Indiana Clean Manufacturing Technology and Safe Materials Institute (CMTI), located at Purdue University.   The institute staffed, performed, and managed the training and evaluation portions of the grant.  The objectives of the Indiana spray training program are four fold: 

         1.  Improve the economic viability of the FRP consortium members by reducing wasted overspray, thereby, reducing raw material costs

         2.  Reduce emissions of styrene to the environment

         3.  Reduce worker exposure to the hazardous materials used in the production process

         4.  Improve the quality of the open mold-produced parts through more consistent spray operations

          This paper reports the findings of the emissions study, which CMTI performed from March through September 1999, to evaluate the near- and long-term effectiveness of the spray training program.

          In order to ensure real world applicability of the training program's effectiveness, an emission analysis was performed at Viking Formed Products, now known as Prodesign Composites (a division of Coachmen  Industries,  Inc., located in Middlebury, Indiana). Prodesign Composites is a leading FRP manufacturer with an admirable history of leadership, innovation, and community involvement.  The emission tests were performed at one of the company's open mold production facilities.  Emission tests were performed under actual production conditions using production molds, parts, gel coats and resins, spray equipment, and Prodesign Composites personnel.  Only one modification was made to the manufacturing plant—the open mold spray operation was totally enclosed to meet EPA Method 204 requirements for total enclosure and 100 percent capture of VOC emissions.


       The open mold FRP industry produces a diverse array of products:  bathtubs, showers, van and truck tops, boats, and a variety of custom products.  Nearly all of the production is based on the spray application of styrene-based gel coats and resin systems.   The gel coat and resin materials are sprayed onto a mold which is the inverse image of the product.  When the material cures and becomes a hard, structurally-sound material, it is separated from the mold.  After further preparation, the molded structure ultimately becomes the finished product (a van top, shower stall, etc.).  The spray process atomizes the gel coat and resin materials, and this atomization contributes significantly to the emission of styrene vapors into the air.

         Styrene emissions have come under increased scrutiny during the past several years due to the categorization of styrene as a hazardous air pollutant (HAP) by the EPA (United States Environmental Protection Agency).  As a result of federal mandates, the EPA is developing a Maximum Achievable Control Technology (MACT) standard for the open mold FRP industry. The MACT will be designed to reduce the industry’s emissions of styrene and other HAPs.  The MACT standard draft is scheduled for promulgation in the spring/summer of 2000.   The draft MACT will become official one year later, and existing FRP companies will probably have two to three years to become compliant with the rules, regulations, and requirements.  Companies will probably be required to use gel coats and resins containing reduced concentrations of styrene (and other HAPs) as well as to implement new application technologies and work practices that reduce emissions of HAPs. A variety of pollution prevention techniques are at the heart of strategies to reduce HAP (styrene) emissions—key among these may be spray-operator training, which CFA emission investigations have shown not only reduces usage, but also reduces HAP emissions.

         In recent years, a new resin application technology, generically referred to as “flow coating,” has been developed.  The new flowcoat technology employs a non-atomizing application process, which has been proven to dramatically reduce resin-born styrene emissions during the application process.  Consequently, most open mold FRP manufacturers currently apply resin with the new non-atomizing flowcoat technology; however, due to the fine-finish surface requirements of the gel coat process, flowcoating technology has not yet proven to be easily adaptable to gel coat application.   Therefore, gel coat application in the FRP industry is still primarily applied via spray atomization.  Investigative work by the CFA has shown that operator-spray training, which instructs spray operators in the proper methods to reduce overspray and reduce (optimize) atomization during the spray process, can reduce styrene emissions to levels commensurate with those of flow-coat application technology.  Furthermore, CFA assessment has shown that operator-spray training has little effect in reducing emissions stemming from the optimal levels achieved with flowcoaters.   Therefore, the operator-spray training program, which became the foundation of the Indiana FRP consortium, concerned itself only with atomized forms of application of gel coat and resin.  Training  for the already optimal flow coat technology was not considered.

         The CFA has developed a complete, operator-spray training program, which is called “Controlled Spray Training.”  The Controlled Spray Training program is a direct outgrowth of the work the CFA performed to develop styrene emission factors for the EPA and for industry.  The CFA experimented with different operator spray techniques and spray gun setup procedures to optimize the process, so that emissions are kept to a minimum during the spray process.

Training Description

         The spray-operator training program is a mixture of classroom lectures, hands-on training in spray equipment setup procedures, operator-technique training, and open mold modification techniques, where applicable.  First, the spray-operator receives a formal, multimedia presentation describing the emission characteristics of styrene-based, open mold production. The operator is then instructed in hands-on, individualized spray training.  The spray-operator is instructed in proper gel coat and resin spray gun setup procedures (typically, the gel and resin coat application process is accomplished utilizing air-assisted airless spray systems).  The gun is initially set up with a minimum of air pressure (30 to 35 psi); the fluid pressure is then adjusted until a tailless, long and narrow ellipse is created.  This new setting is designated the optimized air and fluid pressure setting, and it differs for each gun, depending on the spray tip and the gel coat or resin viscosity.  The spray-operator is instructed to continuously monitor this optimal level throughout the day and change the settings as the situation demands.  When the gun setup optimization is complete, the operator is trained in spray technique.  The Indiana program involves video taping of the individual operator while spraying, prior to receiving any training. The operator is then asked to critique his/her own performance, based on the knowledge gained from the formal, lecture multimedia presentation.  The Indiana training program has found that this self-analysis is extremely effective in quickly changing spray-operator habits.  An added benefit—the instructor is not put in the position of being viewed as overly critical of new, or experienced, spray-operators.  The video taping feedback session gives the instructor and the spray-operator the opportunity to review application technique issues, such as banding, gun angle, spray distance, spray coverage/overlap, etc.

         The CFA Controlled Spray Program also calls for a four to eight inch flange to be placed along the outside edge of the mold.  The flange serves no final product function; it is merely used to catch overspray as the operator applies the coating up to, and slightly over, the edge of the mold.  This reduces the footprint (surface area) of overspray and, consequently, reduces emissions.  The Indiana spray training program recognized that most of the FRP consortium manufacturers used traditional flangeless molds.  Early investigations regarding the adaptability of flanges to existing molds, demonstrated that such adaptation could be costly.  Therefore some companies may decide not to use flanges on existing molds.  Because there are elements of controlled spray, other than mold flanges, which have economic and environmental advantages, the Indiana program involves spray training with and without mold flanges.

Emission Analysis Process Description

         CMTI and CFA staff members visited several open mold FRP manufacturing plants to determine the suitability of each plant to enclose an area so that the area could capture 100 percent of the emissions.  Prodesign Composites proved to have a production facility ideally suited to the task and agreed to act as the emission testing facility.  The company supplied the materials and manpower to completely enclose an open mold spray booth  area.  Prodesign Composites also made changes to the air exchange system so that the booth's make-up air would slightly under-balance the outgoing stack exhaust air.  The under-balance of make-up air allowed the totally enclosed spray booth to operate at a slightly negative pressure, relative to the surrounding nonspray booth production facility.  When the booth enclosure was completed, and the make-up air adjustments were made, the booth met all the requirements of EPA Method 204 Temporary Total Enclosure, the standard for ensuring 100 percent capture of VOCs generated from a production process.  The spray booth's floor area measured approximately 35 feet by 48 feet and the booth's height was approximately 12 feet.

         The emission testing was performed in three phases.  The first test involved a baseline emission analysis.  No spray training was performed prior to the baseline test, and one of the plant's best sprayers (spray operator with many years of experience) served as the test's open mold spray operator for all of the emissions tests.  The second emissions test took place approximately four days after the spray operator received the Indiana FRP consortium's spray training program.  The third emissions test was conducted approximately five months after the second test.  The spray operator received no additional training during the five-month span; however, he did receive verbal mentoring at approximately six-week intervals up until the final test.   The veteran sprayer was selected as the test's applicator because most sprayers in the industry develop expertise only after they have been taught by other seasoned applicators.  The quality of the part produced is strongly influenced by the abilities of the gel and resin sprayer.  If the FRP consortium had selected an inexperienced applicator, the baseline test would have shown very high emissions and poor part quality.  The test was designed to test a typical sprayer's improvement potential, and a typical sprayer is a multi-year veteran.  All emissions tests utilized the company's standard production gel coat and resin as well as standard production air-assisted airless spray gun equipment.  Each test utilized three, identical, standard, van top production molds (78.6 square feet, each).  The parts produced from each of the tests were treated as standard product at the completion of each test.   The mold coating specification was as follows:  24 mils gel coat, 85 wet-mils resin (1st layer) and 85 wet-mils resin (2nd layer).  Woven hand lay-up glass mat as well as oriented strand board materials were also used in the molds.  In addition to the use of the same spray operator for each of the three tests, the same three-person, resin/glass chop roller team was also utilized.

Experiment Design

         The standard spray booth was completely enclosed, and the natural draft area was observed to consistently have incoming air velocities in excess of 200 linear feet, per minute.   New exhaust filters for the booth were weighed and installed and a double layer cover (Kraft paper) was installed over the entire spray booth floor.  The top layer of the Kraft paper floor cover was weighed before and after each test.  The emissions and other pertinent test parameters were monitored as depicted in Figure 1.  All of the mold parts and flange attachments were weighed prior to, and after, the emissions tests.  The spray operator and the three-member resin/glass roller team wore pre-weighed Vortec overalls, booties, gloves, and head socks.

Figure 1














Test 1:  Baseline (no training) Emission Analysis Test

         The spray operator set up the gel coat gun as well as the resin/chop spray gun to his regular spray preference:

       gel coat air-assisted airless settings:  78 psi air to fluid pump, 30 psi catalyst air pressure, .026 tip 40 angle, 4.62 pounds/minute flow, 2% DDM-9 MEKP catalyst

       resin chop air-assisted airless settings:  1100 psi fluid pressure, 40 psi catalyst air pressure, 110 psi chopper pressure, .062 tip 30 angle, 15.5 pounds/minute flow, 1.5% DDM-9 MEKP catalyst, 32% glass

When the spray booth reached a baseline level of less than one part-per-million (ppm) styrene concentration, the spray operator was given a signal to begin the regular spray process.   The operator sprayed the gel coat on each of the three production van top molds.  The molds did not have any flange material attached.  Overspray occurred when the spray operator sprayed the gel coat and resin material, spray gun-to-part distance varied from 20 inches to 3 feet, and gun angle varied from 90 to 45 degrees (referenced to the van top's surface).  The spray operator moved about the mold perimeter only slightly.  The resin/chop spray process began approximately forty-five minutes after the completion of the first mold's gel coat application.  The first layer of the resin was applied to the first mold, then to the second, and then to the third mold.   The resin/chop roller team began the rolling process on the mold when the spray operator finished spraying it.  The spray operator began to apply the second coat of resin to the first mold approximately 30-45 minutes after applying the first coat of resin to the third mold.  This allowed a desired level of polymerization to take place prior to the application of the second resin coat.  The second coat of resin application was performed identically to the first resin coat application (for all three molds), with the exception that the spray operator also sprayed resin onto a mat of woven glass as it lay on the booth floor.  This resin-saturated mat was then hand-placed by the roller team onto specific areas of the mold and rolled down into the wet resin on the mold.  The emission test was complete only after all the molds received their second coat and the booth's styrene emissions had dropped to baseline (less than one ppm).   The baseline emission level was reached approximately 40-55 minutes after the end of the resin application to the third mold.

Test 2:  Emission Analysis Immediately After Spray Training

         The gel coat gun, as well as the resin/chop spray gun, was set up according to the CFA Controlled Spray Training spray gun setup protocol.  The optimum, lowest operating air and fluid pressure was set in order to achieve the optimum, tailless, elliptical spray pattern, with the optimum atomization of the gel coat and resin material:

       gel coat air-assisted airless settings:  57 psi air to fluid pump, 25 psi catalyst air pressure, .021 tip 40 angle, 2.6 pounds/minute flow, 2% DDM-9 MEKP catalyst

       resin chop air-assisted airless settings:  670 psi fluid pressure, 22 psi catalyst air pressure, 66 psi chopper pressure, .052 tip 30 angle, 6.4 pounds/minute flow, 1.5% DDM-9 MEKP catalyst, 32% glass

Each mold was equipped with 4-inch wide flange material attached to the mold edge.  The spray operator sprayed the gel coat and resin material with little overspray.  The operator used the controlled spray training "banding" technique to spray around the part's outer edge in one continuous application, the spray gun-to-part distance varied from 12-24 inches, and the gun angle varied from 90 by plus or minus 20 (referenced to the van top's surface).  All other application procedures were identical to the methods followed in Test 1.

Test 3:  Emission Analysis Five Months After Spray Training Was Conducted

         The gel coat gun, as well as the resin/chop spray gun, was set up to the identical pressures and flow rates as used in Test 2:

       gel coat air-assisted airless settings:  57 psi air to fluid pump, 25 psi catalyst air pressure, .021 tip 40 angle, 2.6 pounds/minute flow, 2% DDM-9 MEKP catalyst;

       resin chop air-assisted airless settings:  670 psi fluid pressure, 22 psi catalyst air pressure, 66 psi chopper pressure, .052 tip 30 angle, 6.4 pounds/minute flow, 1.5% DDM-9 MEKP catalyst, 32% glass

         The molds were equipped with 4-inch wide flange material attached to the mold edges.   The spray operator sprayed the gel coat and resin materials with little visible overspray.  The spray operator used a modified spray training "banding" technique to spray around the part's outer edge in four separate steps.  The operator banded one side edge and then sprayed from the edge to the part's center in a pattern  parallel with the edge.  This pattern was repeated for each of the three remaining edges.  The spray gun-to-part spray distance varied from 12-24 inches, and the gun angle varied from 90, by plus or minus 20 (referenced to the van top's surface).  All other application procedures were identical to the methods followed in Test 1.

Data Acquisition System and Emission Test Quality Assurance

         The FRP consortium's emissions tests were conducted in accordance with the following EPA recommendations:

         < EPA Method 1, Sample and Velocity Traverse for Stationary Sources

         < EPA Method 204, Temporary/Permanent Enclosure—Collection of 100% Emissions

         < EPA Method 25A,  Determination of Total Gaseous, Organic Concentration, Using Flame Ionization Analyzer (FIA or FID [flame ionization detector])

         The equipment employed in the emission tests is listed below:

         < Air-assisted airless gel coat gun; .026 and .021 tip size

         < Air-assisted airless resin chop; .062 and .052 tip size

         < J.U.M. Engineering, Inc.:   flame ionization detector (FID) total carbon analyzer, Model 3-100

         < Dwyer Instrument, Inc.:   2 standard-design pitot tubes, model 160 series

         < Dwyer Instrument, Inc.:   primary standard manometer, model #424

         < NEC Versa data-logging Pentium portable computer

         < National Instruments: LabVIEW, version 5.1 Graphical Programming Software

             data acquisition software

         < National Instruments: LabVIEW DAQCARD AI-16XE-50 voltage to digital


         < National Instruments: SCB-68 voltage to digital interface

         < Dwyer Instrument, Inc.:   pressure transducer, model 607-4—converts stack air

                velocity pressure (in inches of water) to linear voltage readout

         < Alnor Velometer, series 6000:  air velocity measurement instrument

         < Barnant temperature and relative humidity logger, model 6919000

         < Dwyer Instrument, Inc.:   temperature meter-voltage readout, model 4151D

         < Sartorious scale:  360 pounds maximum 2 grams accuracy

         < Sartorious scale:  150 pounds maximum 1 gram accuracy

         < Stack VOC sample line insulated and temperature controlled by a CAL, series 9000 microprocessor

         CMTI staff developed a data acquisition collection system that is capable of simultaneously, and continuously, monitoring pertinent stack data, including velocity, temperature, relative humidity, and parts-per-million of volatile organic compounds (VOCs).  An electronic pressure transducer connected to a stack pitot tube continuously monitored stack air velocity.  Electronic temperature and humidity instruments continuously monitored the stack temperature and relative humidity.  A J.U.M.  Engineering, Inc. total carbon analyzer, flame ionization detector (FID) continuously monitored the parts-per-million of styrene in the stack's exhaust air.   The plant location's barometric pressure (later adjusted to the site's altitude) was obtained from the area's local airport weather data.  All of these data were fed into an especially designed data acquisition software program and readings from each data source were logged every two seconds.  Each of the FRP consortium's three emissions tests spanned an average of 250 minutes.  This culminated in the collection, storage, and computation of more than 30,000 data points for each test.   All test equipment had been calibrated and the equipment's accuracy had been verified.  The J.U.M. FID was calibrated before and after each test, using certified EPA calibration gases (propane standard).

         The stack velocity, in addition to being continuously monitored via the pressure transducer download to the computer, was manually traversed, according to EPA Method 1; therefore, a primary standard was used, in conjunction with the continuous pressure transducer, to develop the approximate dry standard cubic feet, per minute (dscfm ), flow rate of the exhaust stack air.

         The data acquisition software enabled the CMTI staff to calculate the rate of exhaust emission every two seconds; thus, the emissions from the tests were totaled by summing thousands of two-second integration packets.  This proved to be a highly accurate means of collecting and calculating the emission test data.   Figure 2 lists the summary of data collected by CMTI staff during an emission collection efficiency test at the Prodesign Composites facility.  A known amount of pure styrene was atomized into the booth and the emissions were monitored in a manner identical to all of the emission tests.  The capture rate of the known amount of styrene, using CMTI monitoring equipment, exceeded 98 percent.  This on-site collection experiment verified the accuracy and quality of the FRP consortium's emissions testing program.

Figure 2

Emission Collection Efficiency


Styrene Dispensed (lbs.)


Styrene Emissions Collected (lbs.)


Collection Efficiency (%)


Span Gas Drift Check


Styrene emissions, detected in the exhaust stack via the total carbon analyzer, were recorded, in parts-per-million, as propane.  Styrene is an unstable, reactive chemical both in liquid and vapor form.  Thus, the total carbon analyzer, also called a flame ionization detector (FID), for the FRP consortium tests was calibrated with EPA-certified calibration gases composed of known concentrations of propane gas.  Propane was used as the calibration gas because, unlike styrene, it is an extremely stable gas.  Therefore, pure propane gas blended with a diluent gas, such as air or nitrogen, to specific, known concentration will remain at the blended concentration for months.  The stable, blended, and known propane concentration is then used to calibrate the total carbon analyzer and to monitor the instrument's consistency, over time, in accurately recording ppm in an exhaust stack air stream.  The FID is calibrated to sense the carbon atoms in each propane molecules. The instrument assumes that the source of all detected carbon is propane.  Consequently, when the instrument senses the eight carbon atoms from each styrene molecule, it assumes it is detecting 2.667 (83) propane molecules.  Thus, the ppm readout of the instrument for detecting styrene, when calibrated using propane calibration gases is 2.667 times higher (theoretically) than the styrene's actual concentration.  Therefore, to convert from the propane ppm calibrated readout to the styrene ppm, the propane ppm value must be multiplied by a conversion factor of 0.375 (38).  However, since FID instruments have individual sensitivity/response profiles, the actual conversion factor is rarely equal to the theoretical value.  In the case of the FID instrument used for the FRP emission tests, the actual propane-to-styrene conversion factor, determined from multiple testing of freshly prepared known styrene samples, was approximately 0.4248.  This is a typical number and is in agreement with a general range reported by other researchers performing similar tests and using similar equipment.

Figure 3







Test 1 




Test 2

(Controlled Spray)



Test 3 

(Follow-up with controlled spray)                              


      Gel coat phase <+2%
        Resin phase +6.1%

* span gas drift of 5%, or less, from initial calibration to the end of the test is deemed acceptable

         Just prior to the beginning of each test, the FID was calibrated using the EPA-certified propane-standard gases.  At the completion of each test, the FID was rechecked with the same EPA-certified calibration gases.   Figure 3, the FID Calibration Drift Check, reports the difference in the reading of the known ppm EPA gases from initial calibration at the beginning of each test to the end of each test.  The data demonstrate that the FID operated effectively and met the standards of the EPA Method 25A for resistance to drift.

         Figure 4 lists the average, observed actual stack flow rate in dry standard cubic feet, per minute, of each of the three tests.  The maximum variance (from test to test), from the combined average of the three tests, was less than one percent (0.90%).  This result demonstrates the consistent stability of the exhaust system and the stack flow rate monitoring system.

Figure 4


                                                                                                                      Individual Test
                     TEST                                   Average Flow Rate                   Deviation from
               Identification                                  of Each Test                         Combined Tests
                                                                              ft3/min.                            Average

Test 1
(Without controlled spray)                                 23,924.3                                    0.90%

Test 2
(With controlled spray)                                      23,677.7                                    0.14%

Test 3
Follow-up (with controlled spray)                      23,529.1                                    0.76%

Test Results and Discussion

         The CFA Unified Emission Factor Model (UEF) (the emission factor referenced in Indiana air permits) lists gel coats, applied with uncontrolled spray with less than 33% styrene content (by weight), as having an emission factor of 44.5% of the available styrene.   The gel coat and resin used in the FRP consortium's emissions test had a styrene content (by weight) of 27.25% and 33.5%, respectively, (no methyl methacrylate, or any other VOC [other than the minimal catalyst content], was present in either the gel coat or the resin) and the observed uncontrolled spray emission factor was 41.3% available styrene (Figure 5).  Thus, the emissions factor observed in the FRP consortium styrene emissions test is 7.2% less than the CFA UEF model estimate.

Figure 5

Gel Coat Spray Tests 1, 2, and 3

                                                                                            Test 3
                          Test 1            Test 2                                    5-month
                          without          with                                       follow-up
                          controlled      controlled     % difference       with controlled      % difference
                          spray             spray           from Test 1        spray                    from Test 1

Weight of
gel coat
sprayed              49.42 lbs.      40.05 lbs.         -19%                 33.05 lbs.               -33.1%

% available
by weight,
in gel coat)         41.3%           39.2%              -5.1%                40.8%                  -1.2%

savings from
use reduction      NA               $3.10, per          NA                  $5.45, per                NA
                                               mold                                          mold


         The CFA's UEF emission factor for controlled spray of gel coats is listed as 32.5% of the available styrene.  This is a 27% reduction when compared to the uncontrolled spray of gel coat in the UEF.  The FRP consortium's tests demonstrated a maximum of 5.1% (second test) and a minimum of a 1.2% (third test) emission reduction for controlled spray applications of gel coat when compared to uncontrolled spray.  The reasons for the difference in the reduction estimates of the UEF and the FRP consortium's tests might be explained by the following:

         1.   The mold type used in the FRP consortium's test was female while the mold type in the CFA tests (which developed the UEF) was male.  Male molds, due to their geometric configuration, will have substantially more overspray potential than a female mold counterpart.

         2.   If controlled spray techniques are practiced while spraying onto a mold's surface, much less overspray emission reduction is achieved with a female mold than with a male mold.

         3.   The perimeter to surface area ratio was greater for the CFA mold than for the FRP consortium's mold (0.835 to 0.512, respectively); therefore, there was less opportunity for overspray, during overall spray time, with the FRP consortium's mold when compared to the CFA's mold (i.e., overspray of the FRP consortium's mold was less, as a percent of the total amount sprayed, when compared to the CFA's mold).

         4.   The CFA sprayed the same weight of gel coat in its uncontrolled spray tests as it did in its controlled tests; thus, the thickness of its controlled spray parts was greater.  The FRP consortium's tests applied the same thickness to both the uncontrolled spray and controlled spray parts and, thus, reduced the amount sprayed. As previously indicated, surface area is the "after application" determiner of emissions: the thicker a part, given a constant surface area, the less the percentage of emissions per pound applied.  Therefore, the CFA emission factor for controlled spray predicts a greater reduction in percent emissions because more material was put on the part (it was thicker).  The percentage of emissions for the thicker part is lower, and this increases the percent emissions reduction.  The FRP consortium's molds did not have part thickness increase as a result of the controlled spray process.  The part thickness for Test 2, when compared to Test 1, was nearly identical (average, 24 mils), while the part thickness for Test 3 was actually, on average, 22 mils (averaging 2 mils thinner than Test 1's parts).  The fact that 20% less material was sprayed in Test 2 and 33% less material was sprayed in Test 3 means less material sprayed per given surface area (particularly, for Test 3), and this would lead to a higher percent emission factor (compared to the CFA emission estimates) simply due to the thickness (mass) to surface area phenomenon which determines emissions.

         The raw material savings, resulting from the reduction in consumption, is of particular note.  Test 2 demonstrated that controlled spray could reduce gel coat material use by 19%, resulting in a $3.10, per mold, savings.  Test 3 demonstrated an even greater use reduction of 33%, resulting in a $5.45 savings, per mold.  As stated earlier, review of mold thickness data demonstrated that the Test 3 average thickness was slightly less than that of Tests 1 and 2, so conservative cost savings estimates default to Test 2's results.

         The Figure 6 chart illustrates the styrene emissions level (ppm) measured during the gel coat spray phase of Tests 1, 2, and 3.  Thus, the chart compares the concentration of styrene vapors of gel coat spraying without controlled spray training (Test 1), with controlled spray training (Test 2), and measuring controlled spray training information retention (Test 3 follow-up).


         The reduced concentration of styrene ppm observed in Test 2 clearly demonstrates that spray training is effective in reducing styrene emissions and reducing peak ppm levels (see Figure 6 for peak ppm levels attained).  Test 3, performed five months after the spray training program instruction, identifies two additional, important points:

         1.   The effectiveness of the earlier spray training is obvious—the data suggest the sprayer retained about 25% of the skills attained during training.

         2.   The gel coat application speed was significantly slower in Test 2 than in Test 1; however, as the sprayer adapted the training information and technique to the production-oriented routine, the gel coat application speed of Test 3 increased to more nearly match that of   Test 1 (see Figure 6 duration of test).

         While the data from one series of tests (using one spray gun operator) are limited in statistically inference potential, the information suggests that if the training were repeated after an interval of, perhaps, three months, sprayers might retain a greater percentage of training information and sustain the desired technique, thus, more effectively maintaining emission reduction effectiveness.

         The sprayer's ability to increase the "trained" spray application speed nearer to the production rate of "untrained" spray application demonstrates that sprayers can adopt low emission spray techniques and still maintain production rates.

         Figure 7 lists important overspray data, which were collected during Tests 1, 2, and 3.

Figure 7

Overspray Data

                                                                                            Test 3
                          Test 1            Test 2                                    5-month
                          without          with                                       follow-up
                          controlled      controlled     % difference       with controlled      % difference
                          spray             spray           from Test 1        spray                    from Test 1

Weight of                                                                                                                                
on booth
floor                   18.9 lbs.           1.9lbs.            -90%                    .4 lbs.                   -98%

Weight of
overspray on
booth filters            .8 lbs.            .5lbs.             -36%                    NA                       NA


         The value of controlled spray is demonstrated in the table above (Fig. 7).  The weight of the material sprayed past the mold and deposited onto the floor during Test 1 (without controlled spray) was reduced by 90% during Test 2, which utilized a controlled spray technique.  The gains in limiting overspray were more graphic in the controlled spray follow-up test, which demonstrated a 98% reduction in overspray to the floor, when compared to the first test.

         Reduction in overspray to the booth filters also attests to the effectiveness of controlled spray.  The overspray deposited onto the filters during Test 1 (without controlled spray) was reduced by 36% during Test 2, which utilized the controlled spray technique.  Due to a recording error, the booth filter weight change of the controlled spray follow-up test could not be calculated.

         Figure 8 lists the important data collected for the resin spray application for emission tests 1, 2, and 3.

Figure 8

Resin Spray Tests 1, 2, and 3

                                                                                            Test 3
                          Test 1            Test 2                                    5-month
                          without          with                                       follow-up
                          controlled      controlled     % difference       with controlled      % difference
                          spray             spray           from Test 1        spray                    from Test 1

Weight of
sprayed              254.9 lbs.      246.7 lbs.          -3.2%              221.3 lbs.              -13.2%

% available
by weight,
in resin)               16.3%           12.6%              -22.4%            18.9%                    +16.4%

savings from
use reduction         NA            $2.78, per          -3.2%              $11.20, per            -13.2%
                                               mold                                          mold

% glass
(% of total)           27%*         31%*                +14.8                  35.6*                    +32%


         *     Target glass content for the part produced was 30-32%

               A serious malfunction in the resin spray equipment cast serious doubt about the reliability and applicability of the resin emission data obtained from Test 3; however, despite the equipment malfunction, the test was continued and parts were produced.

         The amount of resin used to produce the parts for Test 1 (uncontrolled spray) and Test 2 (controlled spray) remained relatively constant, varying by only 3.2 percent; however, there was a substantial difference in emissions between the two tests (16.3% versus 12.6%):  the controlled spray (Test 2) registered a 22.4% reduction in emissions when compared to the uncontrolled spray (Test 1).  Serious equipment malfunction occurred during the resin coating phase of Test 3, which casts serious doubt on the applicability of the resin emission data gathered for that test.  Despite repeated attempts to correct the malfunction, the pump/spray equipment continued to operate poorly.  Repeated glass chopper jamming also occurred, which caused numerous spray process "starts" and "stops."  Emission test team members experienced eye irritation and also noticed droplet deposits on note pads, cameras, and on their persons.  It was decided that, since this was the FRP consortium's only opportunity for an analysis of resin spray training's long-term effectiveness, the test would continue, the data would be collected, and parts would be produced.

         Test 3 FID calibration/drift check (Figure 3) references a drift of <+ .02% for the gel coat phase and +6.1% for the resin coating phase.  The 6.1% drift is significantly higher than the more typical drift of less than +2%.  The FID drift means that it detected 6% more propane at the end of Test 3 that it had detected when calibrated at the beginning of the test.  One can therefore assume that the FID would detect styrene concentration at a level up to 6% higher than actual concentrations.  If one assumes a worst case condition, that the 6% "over-detection" started at the beginning of Test 3 and continued to the test's completion, then the styrene emissions during the resin application phase would have been overstated by 6% for the entire test. Therefore, the previously mentioned complications regarding Test 3 resin application emissions (due to equipment malfunction) were further exacerbated by the FID's 6% drift from the initial calibration.  If the overstatement assumption is adopted, then the initial emission factor developed for Test 3 (18.9% [Figure 8]) is 6% higher and, in actuality, the emission factor should be reported as 17.8%.

         Even considering the FID-adjusted emission factor of 17.8%, the Test 3 (follow-up trained spray) percent emissions exceed that of Test 1 (no spray training).  While the abnormal FID drift made Test 3's emissions appear higher than actual emissions, the real problem involving the resin phase emission test was, primarily, due to equipment malfunction.  The reader is advised to keep in mind that Test 3's resin emission data are flawed, due to the equipment malfunction, and the test is not representative of what might be expected from typical, follow-up, controlled spray conditions.

         The CFA UEF factor for uncontrolled spray of resin material containing 33.5% is 17.7% available styrene.  Test 1's emission factor was 16.3% available styrene and comports well with the CFA's conservative UEF factor.  The Test 2 (controlled spray) emission factor of 12.6% represents a 22.4% reduction in emissions when compared to that of the uncontrolled spray test.  This represents a substantial reduction in emissions, despite the fact that only 3.3% less material was used.  It follows that controlled spray technique, per se, contributes significantly to emissions reduction beyond that reduction achieved simply from the use of less material.  The Test 1 (uncontrolled) cost savings data when compared to that of Test 2 (controlled spray) suggest that $2.78 per mold can be saved in resin raw material.

         Figure 9 charts the styrene emissions (ppm) from the resin spray process phase of Tests 1, 2, and 3.  Again, one can easily discern the significant reduction of styrene emissions of Test 2 (with controlled spray training) when compared to Test 1 (without controlled spray training).  One can also observe that Test 2's application speed is significantly slower than that of Test 1.  Due to serious and continuous malfunctioning of the spray equipment during Test 3, only limited conclusions can be drawn from the resin spray portion of that test.


Figure 9



         The FRP consortium's spray training program was initiated to accomplish four, key objectives:

         1.   Improve the economic performance of the FRP consortium members by reducing wasted overspray, thereby, reducing raw material costs

         2.   Reduce emissions of styrene to the environment

         3.   Reduce worker exposure to the harmful raw materials used in the production process

         4.   Improve the quality of open mold-produced parts through more consistent spray operations

         The training effectiveness tests demonstrated that all four objectives can be achieved.

Cleanup Cost Savings

         Cost reductions result from a cleaner working spray booth.   Review of Figure 7 suggests that a facility which employs controlled spray techniques may achieve substantial cleaning/labor savings as well as reduced special and/or hazardous waste disposal costs.  A conservative estimate of potential labor and raw material savings follows.

         Potential Cost savings (per booth):

         Reduced cleanup labor costs 1 hr/day x 5 days/week @ $8/hour   ...........     $     40/week

         Booth cleaning material cost reductions  ......................................................  $    25/week

          Gel coat raw material savings @ $3.10/mold
               (assume 12 molds produced, per day, 5 days/week)  .......................         $  186/week

         Resin raw material cost savings $2.78
               (assume 12 molds produced, per day, 5 days/week)  .......................         $ 167/week

         Yearly potential savings @ 50 weeks, per year, per booth  .........................     $20,890/year

Raw Material Consumption Reduction Savings:

         Reduced raw material consumption, made possible by the spray training, provides material cost savings. The observed raw material use reduction, as well as the lower emission factors from the spray training program, are explained in the following emission calculations:

         Gel coat

         49.42 lbs x 27.25% styrene x .413% =    5.56 lbs. of styrene emitted before spray training

         (49.42 lbs x 19% material use reduction[1]) x 27.25% styrene x 41.3% =1.05 lbs. of styrene not emitted due to use reduction

         40.05 lbs x 27.25% styrene x 39.2% =    4.28 lbs. of styrene emissions after spray training.

         5.56 lbs – 4.28 lbs =   1.28 lbs. total emission reduction due to training and use reduction

               The total emission reduction is greater than the reduction attributable solely to reduced use of raw material (1.28 lbs. - 1.05 lbs. = 0.23 lbs).  The conclusion is that 18% (0.23 1.28) of the gel coat emission reduction is due to spray technique, and 82%  (1.05  1.28) of the reduction is due to material use reduction.  This is graphically depicted in Figures 10 and 11.

Figure 10


Figure 11

Gel Coat Emissions Data


Lbs. Applied (Styrene)

UEF Lbs. Estimated Emissions (Styrene)

Lbs. Actual Emissions (Styrene)

Lbs. Emissions Reduction Due to Use Reduction (Styrene)

Lbs. Emissions Reduction Due to Spray Training Technique (Styrene)

Test 1






Test 2







         254.9 lbs x 33.5% styrene x 16.3% =   13.92 lbs. of styrene emitted before spray training

         (254.9 lbs x 3.2% material use reduction[2]) x 33.5% x 16.3% =       0.45 lbs. of styrene not emitted due to use reduction

         246.7 lbs x 33.5% x 12.6% =   10.41 lbs. of styrene emissions after spray training

         13.92 lbs – 10.41 lbs =  3.51 lbs. emission reduction due to training and use reduction

               The total emission reduction is greater than the reduction attributable solely to reduced use of raw material (3.51 lbs. - 0.45 lbs. = 3.06 lbs.).   The conclusion is that 87.2% (3.06 3.51) of the resin emission reduction is due to spray technique, and 12.8% (0.45 3.51) of the reduction is due to material use-reduction.  This is graphically depicted in Figure 12 and 13.   Note that since approximately 7.5 times [(254.9 lbs x 33.5% [Figure 8] – (49.42 lbs x 27.25% [Figure 5])] more styrene is sprayed during the resin phase than during the gel coat phase, the emission reduction attributable to spray training technique for resin accounts for 87.2% of the reduction when compared to the gel coat's emission reduction of  18 percent.

         Total gel coat and resin emission reduction:

         1.28 lbs + 3.51 =  4.79 lbs. reduction due to training, which includes the material  use reduction

               Therefore, comparing the spray training emissions test (Test 2) to the emissions from the no-training test (Test 1), one will observe a 24.8% total emissions reduction (combined gel coat and resin emissions).

Reduced Worker Exposure:

         In addition to raw material consumption reduction, spray training reduces emissions and worker exposure to hazardous chemicals. Review of Figures 6 and 9 demonstrates the value of controlled spray training techniques in reducing the concentration (ppm) of harmful chemicals in the workplace.   Employees are, therefore, exposed to reduced levels of hazardous compounds and the workplace is safer.

         Figure 6 (gel coat spray) demonstrates that the peak styrene level in Test 1 (untrained personnel) is approximately 77 ppm (lasting less than 20 seconds), while the peak level in Test 2 (trained personnel) and Test 3 (follow-up of trained personnel) are approximately 33 and 50 ppm, respectively.  These data represent a peak ppm styrene exposure level reduction for Test 2, which is 57.1% less than Test 1.  The peak level reduction of Test 3, when compared to Test 1, is 35 percent.

         Figure 9 illustrates that the peak level of styrene emissions from the resin spray in Test 1 is approximately 82 ppm; Test 2 and Test 3 display peak styrene levels of 33 and 65 ppm, respectively;  thus, peak exposure levels are reduced by 59.7% (Test 2) and 20% (Test 3) when compared to Test 1.  As stated previously, the equipment malfunctions, which plagued Test 3's resin spray phase, cast doubt upon the reliability of the resin emission test results.

         The reduced exposure levels achieved through the spray training program may also allow a company to reduce air flow and, thus, create savings in heating energy costs.  A cost savings estimate cannot be developed from the data gathered in this test—but, it is a cost savings which could amount to a substantial dollar figure, especially for facilities with more than one spray booth.  Be aware that adequate safety factor margins must be maintained, if reductions are made, to assure worker exposure  (e.g. OSHA) regulatory compliance.

Figure 12


Figure 13

Resin Emissions Data


Lbs. Applied (Styrene)

UEF Lbs. Estimated Emissions (Styrene)

Lbs. Actual Emissions (Styrene)

Lbs. Emissions Reduction Due to Use Reduction (Styrene)

Lbs. Emissions Reduction Due to Spray Training Technique (Styrene)


Test 1







Test 2






         The emissions and use-reduction data results of the spray training emissions test demonstrate that CFA-based spray training reduces emissions to the environment, reduces employees' exposure to hazardous chemicals, and can lead to substantial cost savings for companies that employ the training.

         We thank the following for their contributions to the success of the program:

-  The Indiana Department of Commerce, that provided the grant which made the training and the emissions tests possible;

-  Robert Lacovara, technical director of the Composites Fabricators Association, located in Arlington, Virginia,  who provided unwavering support and guidance;

-  Robert A. Haberlein, Ph.D., Engineering Environmental Consulting Services, Annapolis, Maryland, whose technical expertise in stack testing procedure helped ensure the program's quality; and

-  Thaddeus J. Godish, Ph.D., of Ball State University whose knowledge, support, and donation of equipment made the emissions testing possible.

         We especially thank the management and employees of Prodesign Composites of Middlebury, Indiana, whose patience, professionalism, and dedication contributed greatly to the success of the emissions testing program.

[1] Referenced from Figure 5

[2] Referenced from Figure 8