The primary hypothesis of this research is that "the lowest exposure results when the worker is to the side of the object being painted (90-deg orientation), with the spray gun in the downstream hand; rather than standing in front of the workpiece as is typically the case (180-deg orientation)." The results of this research clearly illustrate that for the situation examined, (i.e., a flat plate in a cross draft booth), and for most compressed air-spray painting guns, this is not true. Exactly the opposite is the case. The lower exposure will be in the 180-deg orientation. It is only when the source momentum is extremely low that the 90-deg orientation is superior. The effects of motion and in which hand the spray gun is held were examined in wind- tunnel studies. The results suggest that motion seems to produce results midway between the two stationary positions (i.e., 90- deg and 180-deg) and that in the 90-deg position when the spray gun is held in the upstream hand lower breathing zone concentrations result than when the gun is in the downstream hand. However the results of the motion study are not reported with confidence. The spraying motion that the robot-mannequin produced was not a reasonable representation of a real spray painter, as it was quite exaggerated. In addition the robot-mannequin was too large for the small wind tunnel and the data that resulted would not be directly scalable to a real person. In contrast the smaller mannequin which provided good scaling data. The effect associated with which hand the gun is in was supported by 2-d numerical work; but also needs further evaluation. A surprisingly accurate model for estimating transfer efficiency is developed based on a simple impactor theory and size distribution data predicted using the Kim and Marshall reference. This model is confirmed with numerical simulations and is important in estimating overspray generation rates. This model essentially completes the exposure model allowing a dimensionless breathing zone concentration to be predicted as a function of the momentum flux ratio, and geometry. The major positive results may be summarized as follows: (1) Through the use of dimensional analysis and scale model wind-tunnel studies a simple mathematical model was developed to predict human exposure to total mass generated during spray painting operations in cross-draft booths. This model was evaluated very favorably with field studies. (2) Commercially available Computational Fluid Dynamics (CFD) software in conjunction with algorithms developed in-house, were used to generate estimates of human exposure to aerosols generated during spraying operations. These simulations produced breathing zone concentrations that were in good agreement with the field- validated model. (3) Use of a dimensionless breathing-zone concentration as a function of a momentum flux ratio is a powerful tool to assess control intervention alternatives for human exposures to airborne contaminants. There are two important limitations in the study. First is the inability of the commercial CFD package (FIDAP v 8.0) to provide the aerosol concentrations directly. This is because the residence time of the aerosol trajectories within the breathing zone cannot be calculated directly in commercial software (FIDAP v8.01). The information for each trajectory must be written to disk, the resulting file reformatted (using a pert script we developed in-house specifically for the FIDAP output files) and then run through our aerosol algorithm to generate estimates of concentration, transfer efficiency, overspray generation rate, and dermal deposition rates. The fact that these data files approach 1 gigabyte each and over 50 such files may be processed per simulation results in a great inefficiency. Discussions with the FIDAP (FLUENT) software group indicated that access to the source code was not an option, and that developing the capability to do such calculations was not a high priority in upcoming versions of the code. Second, the work here uses, with the exception of the field study, a nonvolatile oil. The effects of evaporation on the model are only indirectly dealt with. While the assumption that vapor will be transported similarly to the aerosol seems a reasonable first approximation, there is clearly room for improvement. The use of total mass (i.e., both vapor and solid) as the dependent variable in the model obscures some of the details that would be useful to hygienists. The results of this research are useful in several ways. The mathematical model of exposure to contaminants generated in spray painting operations allows the industrial hygiene engineer to estimate the effects of changing various parameters on the worker's ! exposure. The model addresses not only the effects of ventilation, but work practices (specifically orientation and distance of spray gun to work piece surface), and transfer efficiency (an hence contaminant generation rate). The computational fluid dynamics approach provides a framework for examining much more complex (and realistic) problems than the simple dimensional analysis model. More complex shapes, downdraft configurations, and even motion are possible with this tool. Useful" numerical flow visualizations" can be produced on rather coarse meshes that provide helpful information on the design of ventilation systems, and for EPA Pre-Manufacture Notification exposure estimations. It should be noted that the general approach of using CFD for the spray painting problem that is outlined here can serve as a blueprint for other processes as well. On a very practical level the work done here shows the importance of work-practices in conjunction with ventilation as an exposure control. The orientation of the spraying with respect to the airflow direction is a critical variable in achieving lower exposure. In addition, this research provides a methodology for developing further, the use of computational fluid dynamics as a tool for occupational hygienists. This is accomplished by a controlled analysis of the uncertatinties involved. The creation of a simple conceptual model based on dimensional analysis provides the data necessary to examine both the CFD results and reality. This allows assessment of the major sources of error and to identify how further improvements can be made efficiently.
Michael R Flynn, ScD, Department of Environmental Sciences and Engineering, University of North Carolina, Chapel Hill, NC 27599-7400
Department of Environmental Sciences and Engineering, University of North Carolina, Chapel Hill, North Carolina