Modeling of Oil Mist and Oil Vapor Concentration in the Shale Shaker Area on Offshore Drilling Installations

Abstract

The objective of this study was to develop regression models to predict concentrations of oil mist and oil vapor in the workplace atmosphere in the shale shaker area of offshore drilling installations. Collection of monitoring reports of oil mist and oil vapor in the mud handling areas of offshore drilling installations was done during visits to eight oil companies and five drilling contractors. A questionnaire was sent to the rig owners requesting information about technical design of the shaker area. Linear mixed-effects models were developed using concentration of oil mist or oil vapor measured by stationary sampling as dependent variables, drilling installation as random effect, and potential determinants related to process technical parameters and technical design of the shale shaker area as fixed effects. The dataset comprised stationary measurements of oil mist (n = 464) and oil vapor (n = 462) from the period 1998 to 2004. The arithmetic mean concentrations of oil mist and oil vapor were 3.89 mg/m ³ and 39.7 mg/m ³ , respectively. The air concentration models including significant determinants such as viscosity of base oil, mud temperature, well section, type of rig, localization of shaker, mechanical air supply, air grids in outer wall, air curtain in front of shakers, and season explained 35% and 17% of the total variance in oil vapor and oil mist, respectively. The developed models could be used to indicate what impact differences in technical design and changes in process parameters have on air concentrations of oil mist and oil vapor. Thus, the models will be helpful in planning control measures to reduce the potential for occupational exposure.

Keywords:

• air concentrations
• determinants
• offshore
• oil drilling
• oil mist
• oil vapor

ACKNOWLEDGMENTS

The authors thank the group of occupational hygienists in the Norwegian Oil Industry Association (OLF) and, in particular, their leader, Vegard Peikli, for valuable advice and discussions during this work. We are grateful to OLF for funding the study.

Notes

^A N = number of rigs with respective values (0 or 1) for the potential determinants.

^B n = number of measurements.

**p ≤ 0.01;

*p ≤ 0.05.

**, significant at P≤ 0.01;

*, significant at P≤ 0.05; otherwise P ≤ 0.20; wrS², within-rig variance; brS², between-rig variance.