Annals of Occupational Hygiene Advance Access originally published online on September 1, 2005
Annals of Occupational Hygiene 2006 50(2):109-122; doi:10.1093/annhyg/mei049
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© 2005 British Occupational Hygiene Society Published by Oxford University Press
Original Article |
Exposure to Oil Mist and Oil Vapour During Offshore Drilling in Norway, 1979–2004
Department of Public Health and Primary Health Care, Section for Occupational Medicine, University of Bergen, Kalfarveien 31, N-5018 Bergen, Norway
* Author to whom correspondence should be addressed. Tel: +47-55-58-61-57; fax: +47-55-58-61-05; e-mail: kjersti.steinsvag{at}isf.uib.no
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Objectives: To describe personal exposure to airborne hydrocarbon contaminants (oil mist and oil vapour) from 1979 to 2004 in the mud-handling areas of offshore drilling facilities operating on the Norwegian continental shelf when drilling with oil-based muds.
Methods: Qualitative and quantitative information was gathered during visits to companies involved in offshore oil and gas production in Norway. Monitoring reports on oil mist and oil vapour exposure covered 37 drilling facilities. Exposure data were analysed using descriptive statistics and by constructing linear mixed-effects models.
Results: Samples had been taken during the use of three generations of hydrocarbon base oils, namely diesel oils (1979–1984), low-aromatic mineral oils (1985–1997) and non-aromatic mineral oils (1998–2004). Sampling done before 1984 showed high exposure to diesel vapour (arithmetic mean, AM = 1217 mg m–3). When low-aromatic mineral oils were used, the exposure to oil mist and oil vapour was 4.3 and 36 mg m–3, and the respective AMs for non-aromatic mineral oils were reduced to 0.54 and 16 mg m–3. Downward time trends were indicated for both oil mist (6% per year) and oil vapour (8% per year) when the year of monitoring was introduced as a fixed effect in a linear mixed-effects model analysis. Rig type, technical control measures and mud temperature significantly determined exposure to oil mist. Rig type, type of base oil, viscosity of the base oil, work area, mud temperature and season significantly determined exposure to oil vapour. Major decreases in variability were found for the between-rig components.
Conclusions: Exposure to oil mist and oil vapour declined over time in the mud-handling areas of offshore drilling facilities. Exposure levels were associated with rig type, mud temperature, technical control measures, base oil, viscosity of the base oil, work area and season.
Keywords: exposure • offshore • oil drilling • oil mist • oil vapour
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Offshore exploration drilling on the Norwegian continental shelf started in 1966. Non-Norwegian multinational companies initially dominated the exploration and development of the oil and gas fields, whereas national oil companies became important operators from 1972. The number of exploration wells completed each year increased until 1986, reaching 51. Then the number levelled off and decreased to 23 in 2003. Until 2004, a total of 1063 exploration wells had been completed (Ministry of Petroleum and Energy, 2004
). Drilling of development wells increased gradually to 
190 in 2001. A total of 2329 development wells were started from 1973 to 2003 (Norwegian Petroleum Directorate, 2004).
Over the years, the oil companies have engaged many national and foreign contractors specialized in drilling operations. Workers in the drilling crews may be exposed to drilling mud, either by inhaling aerosols and vapour or by skin contact (Davidson et al., 1988). The drilling mud is used for many purposes such as lubricating and cooling the drill stem and bit, providing pressure support in the well and transporting cuttings to the surface (Fig. 1). The fluid is a complex mixture of either water- or oil-based fluids and a large number of additives, depending on the system used (Hudgins, 1991). The water-based system is often used in the upper sections of a well, whereas oil-based mud is the only option in long or deep wells. The composition of these mud systems has varied considerably both in time and between suppliers (HSE, 2000). A typical oil-based drilling fluid used on the British sector of the North Sea comprises (by volume) 52% base oil, 30% water and additives such as weighting materials (11%), emulsifiers (3%), brines (2%), pH increasers (1%) and viscosifiers (1%) (HSE, 2000). The original oil-based drilling muds contained diesel as the base oil (Davidson et al., 1988). Diesel was phased out in the early 1980s and gradually replaced by petroleum-based oils with a reduced aromatic content (HSE, 2000).
|
Using oil-based mud systems may generate airborne hydrocarbon contaminants (oil mist and oil vapour) in the mud-handling areas (Davidson et al., 1988
). The Norwegian Oil Industry Association (1996) assumes the potential exists for inhaling oil mist and oil vapour along the flow line from the top of the well to the separation equipment, which includes shale shakers, desanders, desilters, centrifuges and the mud pits (Fig. 1). They specifically state that cleaning and changing screens on the shale shakers may lead to high exposure levels. Originally these areas were designed for water-based mud and were being kept open, and the control of aerosols and vapour relied on general ventilation.
Under such circumstances, personal exposure to total hydrocarbon compounds has been reported to be up to 450 mg m–3 during work at the shale shakers when drilling with oil-based mud (Davidson et al., 1988). At an installation with a higher level of enclosure of the mud systems, James et al. (2000) reported results from two personal samples in the shale shaker room to be 0.06 and 0.40 mg m–3 for oil mist and 3.2 and 35.0 mg m–3 for oil vapour. Although the offshore oil companies have measured exposure numerous times, published results from this working environment are scarce.
Knowledge about exposure is important in evaluating the health risk for these workers. The current Norwegian 8 h occupational exposure limit (OEL) on oil vapour of 50 mg m–3 (Norwegian Labour Inspection Authority, 2003) is based on a possible increased prevalence of lung fibrosis and lung cancer at exposure levels of 50–100 mg m–3 (Skyberg et al., 1986; Rønneberg and Skyberg, 1988; Rønneberg et al., 1988). The OEL for oil mist (1 mg m–3) (Norwegian Labour Inspection Authority, 2003) is based on irritant effects on the respiratory system at exposure levels of 2.5 mg m–3. However, the OEL for oil mist has to be evaluated also with regards to its content of carcinogens like polycyclic aromatic hydrocarbons (PAHs) (Arnesen, 1985). Due to 12 h shifts offshore, the OELs should be multiplied by 0.6 (Petroleum Safety Authority Norway, 2004). To our knowledge there are no published data on the association between respiratory health effects and inhaled air contaminants from oil-based drilling muds.
The currently used industrial standard for air sampling of oil mist and vapour was developed in 1989, and consists of a series coupling of a glass fibre filter with a charcoal tube backup. This method makes it possible to sample oil mist and vapour simultaneously for 2 h (Malvik and Børresen, 1988; James et al., 2000). Oil mist collected on the filter is analysed by Fourier Transform Infrared Spectrometry (FTIR), while the oil vapour fraction in the charcoal tube is analysed by gas chromatography (GC) with a flame ionizing detector. Although FTIR analysis was recommended for both vapour and mist quantification (Malvik and Børresen, 1988), the industry has preferred GC-analysis for determination of oil vapour. The analysing methods are in accordance with NIOSH-methods 1500 (vapour) and 5026 (mist) (NIOSH, 1994).
Several factors such as mud temperature, mud flow rate, well length, well section and viscosity of the base oil might be expected to influence the exposure levels in the working atmosphere, but the relative contributions of these factors have not been published. Little is known regarding the formation and characteristics of the oil mist produced in the mud-handling areas. Review articles by Eide (1990) and Gardner (2003) suggest that oil droplets are generated by both the vibrating screens of the shale shakers and condensation of vapour, but no studies have been performed to support these theories or to quantify the contribution of these sources of mist.
The objective of this study was to quantify the personal exposure to airborne hydrocarbon contaminants in the form of oil mist and oil vapour in the mud-handling areas of offshore drilling facilities operating on the Norwegian continental shelf when drilling with oil-based muds from February 1979 to May 2004. An additional aim was to identify determinants and describe time trends of exposure by modelling the measurements sampled by the method described by James et al. (2000). This was done as a part of a prospective cohort study on occupational cancer among 28 000 past and present offshore workers, which includes 5000 persons involved in drilling operations (Strand and Andersen, 2001).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Collection of data
Heads of the health and safety departments in 20 companies employing offshore workers were contacted by phone followed by an official request sent by e-mail. Attached to the e-mail was a letter from the Norwegian Oil Industry Association requesting the companies to let a University research group of 2–4 people visit the company to carry out interviews and to collect data on chemical exposure (with particular attention to carcinogens). Visits were made to oil companies (8), drilling companies (5), chemical suppliers (3) and other contractors (4). In addition, one trade union, one employer's association and relevant authorities (3) were visited in a similar manner. All those contacted accepted the visits and interviews. Key informants, selected by the company itself, such as staff responsible for health and safety, offshore drilling personnel, occupational physicians and occupational hygienists were interviewed on the work processes and chemical products used on offshore facilities. A report was written for each visit and returned to the informants for feedback. The reports were then evaluated in cooperation with the main contact in the respective companies.
The research material in this study comprises the final reports from the company visits in addition to monitoring reports and risk assessments, which were made accessible on the visiting day. When companies promised access to more exposure reports, the data collection process continued by 3–15 personal contacts with each company, either through phone (1–4), e-mail (2–12) or additional meetings (0–1). The data collection process took place in 2003 and 2004 and lasted for 18 months.
Database
A database containing information from the monitoring reports was constructed in SPSS 12.0 for Windows (SPSS Inc., Chicago, IL, USA). This database comprised relevant information to characterize exposure to oil mist and vapour in the mud-handling areas. The parameters entered were: rig name, type of rig, purpose of sampling, base oil, base oil characteristics (aromatic content and viscosity), work area, year and month, weather conditions, process parameters (well section, mud temperature, mud flow and well length), sampling and analysis methods, analysing laboratory and occupational hygienist.
Type of rig was divided into fixed or movable drilling facilities (Table 1) according to the practice of the Norwegian Petroleum Directorate (2005). A fixed facility is as a generic term for all facilities placed on a field permanently, whereas a movable facility is not meant to be permanently placed on the field during its lifetime (Norwegian Petroleum Directorate, 2005).
|
The purpose of sampling was divided into whether or not the measurement was performed to determine the effect of technical control measures (Table 1). If information regarding the purpose was not stated (12 samples) or the purpose was to document the effect of changing the base oil (9 samples), the samples were included in the group in which no technical measures had been performed.
Based on information in the material collected and in the literature (Davidson et al., 1988; HSE, 2000), the mud systems used on the Norwegian continental shelf were divided into three generations (Table 2): diesel (1979–1984; aromatic content >15% by volume; boiling point range 150–370°C), low-aromatic mineral oils (1985–1997; aromatic content 1–10%; boiling point range 220–325°C) and non-aromatic mineral base oils (1998–2004; aromatic content <0.01%; boiling point range 230–320°C for normal viscosity oils and boiling point range 210–260°C for low-viscosity oils). When information on the base oil was not stated (Table 2), we assumed that low-aromatic mineral base oils were used from 1989 to 1997, versus non-aromatics from 1998 to 2004. For all samples for which the type of base oil was not stated, it was assumed that the viscosity of the base oil was in the normal range of 3.0–4.5 mm2 s–1 at 40°C (Table 1). Low-viscosity base oils with a viscosity of 2.0–2.3 mm2 s–1 at 40°C are presumed to be more volatile.
|
Occupational hygienists in this industry aim to gather samples when conditions are at the worst case. Drilling at the end of the 12.25 inch section is considered to produce the highest exposure to airborne contaminants because both mud flow and mud temperature are high. Due to fast and unpredictable changes when drilling, departure delay due to bad weather or fully booked helicopters, drilling in the subsequent, narrower 8.5 inch section may occur when occupational hygienists finally reach the platform to do sampling. The mud flow is lower in the 8.5 inch section, leading to less fluid passing the mud-handling area, which is thought to be associated with lower exposure levels.
Detailed data on weather conditions were lacking in many of the reports collected. Splitting the months of the year into summer and winter seasons was therefore selected as a rough indicator of weather conditions.
Data analysis
All exposure data from 1979 to 2004 were stratified by sampling method and base oil, and presented as arithmetic mean (AM), geometric mean (GM) and their respective standard deviations (SD and GSD) (Table 2). The frequency distributions of both oil mist and oil vapour exposure levels were skewed, and the estimated geometric standard deviations were <3 for most of the strata (Table 2). In accordance with Hornung and Reed (1990), the measurements under the limit of detection (LOD) were set as LOD/21/2. Due to the skewed nature of oil mist and oil vapour exposure data, these variables were loge-transformed before statistical analysis (Table 3). Exposure levels in different groups were analysed by t-tests and one-way analysis of variance. Correlations between continuous variables were evaluated using Pearson's correlation coefficient.
|
Categorical variables were dichotomized before analysis (Table 1). Variables included in the exposure models were chosen on the basis of a significance level of P < 0.20 in univariate analysis or on logical assessment of the potential determinants of exposure (Table 1). Linear mixed-effects models were developed to model the time trend and to show the influence of different variables (P to enter <0.05) on personal exposure to oil mist or oil vapour. These models have the same general form as described by Rappaport et al. (2003)
. Since the data were unbalanced, the models were fitted using restricted maximum likelihood estimation. Only 86 of 340 samples had worker identification, reflecting 1–6 repetitions of 16 workers, which was considered to be too few to use worker as a random effect in the mixed models. To account for repeated measurements taken from the same rig, the individual rig was viewed as a random effect. The potential determinants of exposure were set as fixed effects. All statistical analysis and figures were performed using SPSS 12.0 for Windows (SPSS Inc., Chicago, IL, USA).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Personal exposure
Table 2 presents the results from 72 monitoring reports on exposure to oil mist and oil vapour when drilling with oil-based muds from 1979 to 2004. Monitoring reports from before 1989 solely describe sampling of the vapour fraction with either charcoal tubes or dosimeters, and analysis by infrared spectrometry or GC. From 1989 both mist and vapour were collected at the same time on series coupled glass fiber filter and charcoal tube and analysed by FTIR and GC.
In three monitoring reports in 1979, 1982 and 1983, the AM of exposure to diesel vapour (1217 mg m–3) (Table 2) was higher than the Norwegian OEL for white spirits (1050 mg m–3). According to one of the monitoring reports, this OEL was used also for diesel vapour at that time. The reports said that the diesel base oils had an aromatic content of 16–24%.
When low-aromatic base oils (aromatic content 1–10%) were used and sampling was done with filter and charcoal tube in series, the AM exposure was 4.30 mg m–3 to oil mist and 36.3 mg m–3 to oil vapour (Table 2). The percentage of samples exceeding the OEL was 42% for oil mist and 45% for oil vapour. From 1998, when non-aromatic base oils (aromatic content <0.01%) were used, the overall AM exposure to oil mist declined to 0.538 and 16.1 mg m–3 for oil vapour (Table 2), and the percentages of samples exceeding the OEL were 24 and 15%, respectively.
Sampling oil vapour using charcoal tubes alone showed exposure within the ranges of those found when the filter and charcoal method was used (Table 2). Dosimeter sampling of oil vapour showed relatively high exposure to oil vapour compared with the present OEL. The samples collected by charcoal tube alone or by dosimeter (Table 2) were all full-shift measurements (8–12 h) covering different work areas.
Preparation for exposure models
The exposure modelling only included the measurements performed with the currently used filter and charcoal tube sampling method. This led to the exclusion of seven reports.
Descriptive information from the remaining 65 reports covers the period 1989–2004 and includes 22 fixed drilling facilities and 12 movable drilling rigs (Table 3). The number of samples taken at both movable and fixed facilities increased during this time period (Fig. 2). The median number of monitoring reports per rig was 2 (range 1–9), and the median number of samples per rig was 8 (range 1–39) (Table 3). Tables 1 and 4 present the statistical significance levels of the potential determinants of exposure.
|
|
The exposure to oil mist and vapour was significantly higher (Table 1) at the movable units (GM: 0.588 and 20.3 mg m–3) than at fixed facilities (GM: 0.385 and 9.37 mg m–3). The most common purpose of air sampling until 1998 was to document compliance with OELs, without any reference to controlling technical measures (Fig. 3). Later, documentation of technical and physical control measures has become the most important reason for sampling (Fig. 3). Documentation of control measures was associated with lower exposure to oil mist than compliance measurements (Table 1).
|
Most air samples were taken during work in the shale shaker area, where the GM exposure was 0.462 mg m–3 for oil mist and 15.0 mg m–3 for oil vapour (n = 225). The main task for operators in the shale shaker room is to survey the shakers in the shaker room itself or from a cabin. Further, the operators usually sample and analyse mud about four times an hour, change or repair screens on the shakers and clean the room and the shakers. The measured exposure [GM (GSD)] during work in the mud pit area [mist: n = 44, 0.584 (6.63) mg m–3, vapour: n = 41, 10.7 (5.02) mg m–3], the pump room [mist: n = 6, 0.333 (1.46) mg m–3, vapour: n = 6, 15.3 (1.67) mg m–3], the mud laboratory (vapour: n = 4, 16.1 (1.31) mg m–3), the cutting wash room [mist: n = 3, 0.776 (1.83) mg m–3, vapour: n = 4, 59.1 (1.51) mg m–3] and the drilling floor (mist: n = 1, 1.40 mg m–3, vapour: n = 1, 24.0 mg m–3) did not differ significantly from those found in the shaker area (analysis of variance). In exposure models, these areas were merged with the samples from the shaker areas, resulting in overall exposure to oil mist and oil vapour of 0.481 and 14.6 mg m–3 in this group. The exposure to oil mist (0.317 mg m–3) and vapour (4.76 mg m–3) was significantly lower (Table 1) for work in the slurrification unit (n = 53), which is present at some installations. In this area, the cuttings are crushed and blended with water to form a slurry that is reinjected into old wells (Fig. 1). One operator surveys this process, and samples of mud are taken once an hour.
The mud temperature interval covered by the samples was 31–82°C (Table 1). Mud temperature was positively correlated with mudflow and well length, and both mud temperature and mudflow were time-dependent, with increasing trends from 1989 to 2004 (Table 4). Mud temperature was stated for most samples (Table 1), and according to the monitoring reports it has been used as an indicator of worst-case exposure in drilling. In the present study, mud temperature was considered to be a logical determinant of exposure, although in bivariate analysis mud temperature did not indicate the expected positive correlation with exposure (Table 4).
Time trends
In the linear mixed-effects model (Model 0, Table 5) with rig as random variable, the between-rig and the within-rig variance components contributed similarly to the total variance for both exposure parameters. There were significant downward time trends for both oil mist (6.3% per year) and oil vapour (8.4% per year) when the year of monitoring was introduced as a fixed effect in a linear mixed-effects model analysis (Model 1, Table 5). This time trend was mainly associated with a reduction in the between-rig variance for both exposure parameters (Table 5). The observed GM level of exposure for oil vapour and oil mist varied from year to year but seems to have a decreasing time trend from 1989 to 2004 (Fig. 4). However, Fig. 2 shows that most of the data are from the last 6 years and that relatively few samples were taken on movable drilling rigs in the first few years after 1989. In 1992 three reports from one offshore installation showed particularly high exposure levels to oil mist, with all 18 samples exceeding the OEL. One of the reports had 10 extremely high oil mist levels (range 23.4–48.1 mg m–3), which according to the reporting occupational hygienist might be questionable. No evidence of errors made during sampling or analysis was found, and the presence of such high exposures was not excluded. However, when excluding all oil mist results from 1992 (n = 18) from Model 1 (Table 5), the estimated downward time trend is 4.0%.
|
|
Determinant model of exposure to oil mist
Including the significant fixed effects of type of rig, purpose of sampling and mud temperature explained 14% of the total variance in oil mist (Model 2, Table 5). The within-rig variance decreased by 10%, and the between-rig variance declined by 18% compared with Model 0 (Table 5).
For a fixed production rig, this model (Model 2, Table 5) predicts a personal exposure level of 0.41 mg m–3 when drilling with an average mud temperature of 58°C, and the purpose of sampling is to compare with limit values. The categorical variable contributing most to a change in the estimated exposure, by a factor of 2.2, is the type of drilling rig (Model 2, Table 5). On a movable drilling rig under otherwise identical conditions, the exposure level would be 0.90 mg m–3. When the purpose of the air sampling is to document technical changes, the exposure is reduced by a factor of 0.58. The relationship between mud temperature and estimated oil mist exposure on fixed and movable rigs is given in Fig. 5.
|
Determinant model of exposure to oil vapour
In Model 2 (Table 5), the fixed effects of rig type, the type of base oil used, viscosity of the base oil, work area, mud temperature and season explained 27% of the total variance in oil vapour exposure compared with Model 0. The largest decrease was in the between-rig variance (42%), while the within-rig variance declined 11%.
The statistical model (Table 5) suggests oil vapour exposure of 14.6 mg m–3 in the shale shaker area of a fixed installation when drilling with low-aromatic base oil at a mud temperature of 58°C in winter. The exposure is reduced by a factor of 0.45 when drilling with non-aromatic base oil, indicating exposure of 6.6 mg m–3. The respective values for a movable drilling rig are 28 and 13 mg m–3. Low-viscosity base oils were reportedly used in some wells from January 2001. Drilling with these base oils increased exposure by a factor of 2.2. Work at the slurrification units significantly lowered exposure, by a factor of 0.46, whereas drilling in summer increased exposure 1.3 times. Increased mud temperature increases the exposure level. The relationship between mud temperature and estimated oil vapour exposure on fixed and movable rigs is given in Fig. 5.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Time trends
Our results indicate that the exposure to oil mist and vapour has decreased during recent decades, but exposure above the recommended limits is still reported.
The extremely high exposure to diesel vapour in the earliest period of monitoring (1979–1983), which is excluded from the statistical modelling, might be due to the lack of technical control measures in the mud-handling areas. At that time the drilling facilities were designed for water-based mud systems, which were probably expected not to cause harmful health effects. The reduction in exposure from 1979 to 2004 occurred as diesel was being replaced with low-aromatic and later non-aromatic base oils. The boiling point range for the diesel oils includes lower temperatures than the two subsequent generations of base oils. Generally, vapour pressure decreases as the boiling point increases, indicating less evaporation of base oils with higher boiling points. This might partly explain the high oil vapour exposure when diesel base oils were used. Furthermore, since diesel vapour was actively sampled on charcoal tubes during 12 h shifts, it cannot be excluded that some oil mist might also have been collected and resulted in an overestimation of the diesel vapour exposure level.
Technical control measures to reduce exposure have mainly comprised constructing cabins for the operators and installing more efficient ventilation systems. Closing open fluid flow lines and mud pits has probably also made the working environment less contaminated. In addition, the purpose of the air sampling reports has changed through time. Before 1999, sampling was almost exclusively focused on testing compliance with limit values, whereas since then the largest fraction of air samples was to document technical control measures carried out in the mud-handling areas. If the changes were successful, lower exposure would be expected, as indicated for oil mist in the mixed effect model. However, the various types of control measures presumably have different relative impact on the exposure levels. An increased focus from the authorities in the past 7–8 years on documenting the exposure level as an important part of risk assessment might also have initiated the measurement of exposure on newer generations of rigs with lower exposure.
The linear mixed-effects models indicate significant decline over time in exposure to oil mist and vapour from 1989 to 2004 of 6 and 8% per year, respectively. These time trends were mainly associated with decreases in between-rig variance, which might indicate that rigs with lower exposure were included rather than exposure being reduced over time within the respective rigs. This could be explained by the low number of years sampled for most rigs and also few repeated measurements, which were mostly taken in short time frames within the different rigs. Thus, the time trends might partly be functions of the rigs selected for sampling. This data material represents 50% of the fixed drilling facilities and 20% of the movable drilling rigs. We did not evaluate whether these rigs are representative for all the rigs operating in the time period investigated. Further, the time trend should be interpreted cautiously, especially for exposure to oil mist. This time trend seems largely affected by the very high exposure concentrations measured in 1989 and 1992, whereas after 1992 the observed exposure to oil mist seems to be relatively independent of time. However, the magnitude of these time trends is in the same range as those reported for long-term exposure trends in other industries such as the asphalt industry (Burstyn et al., 2000), the carbon black industry (van Tongeren et al., 2000) and the rubber industry (Vermeulen et al., 2000).
Determinant models
The estimated exposure to oil mist and vapour on the movable drilling rigs was about twice as high as on fixed drilling facilities. This can be explained by older technologies with more open flow lines, less developed ventilation systems and the fact that more time was spent in the exposed areas. In addition, these movable facilities are more affected by waves, which lead to more spillage of mud. More details on the design of the mud-handling areas are needed to verify this.
The models indicate that technical control measures prior to sampling have had most effect on oil mist concentrations. While the design of the shakers and mud pits has remained unchanged on most drilling facilities, there have been considerable improvements in ventilation of the mud-handling area on most rigs.
In bivariate analysis, the mud temperature correlated both with mudflow and well length, but none of these parameters correlated unambiguously positively with oil mist or vapour. Most reports stated the mud temperature, and it was therefore chosen as a variable to be entered into the exposure models. The multivariate exposure model agrees with this assumption by indicating that the mud temperature significantly predicts oil mist and vapour exposure, as the exposure increases by 19 and 16%, respectively, for an increase in temperature of 10°C.
The section of the well was not a significant determinant and was not included in the final models.
Exposure to oil vapour was significantly lower for drilling with non-aromatic base oil than with the previously used low-aromatic base oil. Whether this is due to the characteristics of the base oils such as evaporation or to other time-linked changes such as technical control measures or the introduction of newer rigs cannot be determined.
Long and complicated high-temperature and high-pressure wells may require fine-tuned base oils with low viscosity. These low-viscosity base oils have a lower boiling point range and presumably a higher vapour pressure compared to those with normal viscosity. This might explain the increased oil vapour exposure when low-viscosity base oil was used in the present study. Viscosity was not a significant determinant for oil mist, probably because it has little effect on the oil mist produced by mechanical agitation of the shakers.
The workers in the slurrification unit had lower exposure to oil vapour than did workers in the other mud-handling areas. This might partly be because the temperature of the mud was reduced by the time it reached the slurrification unit. The temperature was not measured in these units, so this could not be verified. Few reports stated the actual time spent in the respective work areas and on the specific tasks, and this could not be used for further analysis.
Oil vapour is generated by evaporation from the mud system, especially in the shale shaker area, where solids and liquids separate. Oil mist is presumably produced by a combination of aerosol formation by mechanical agitation of the shale shakers and the condensation of vaporized base oil. Depending on the equilibrium between the vapour and liquid phases, oil vapour produced by evaporation from oil mist might also contribute to the total vapour concentration. One reason for the increased oil vapour exposure during the summer season might be that the higher air temperatures shift the equilibrium between the phases towards increased vapour concentration. Less wind during summer might also contribute to higher exposures. Generation of oil mist appears to be independent of the seasonal effect.
The major decreases in variability in the mixed models were found for the between-rig components. This might be explained by the relatively small ranges of process conditions and the clusters of repeated measurements within a short time frame for most of the individual rigs.
Strengths and limitations of the study
We aimed to get access to as many exposure reports as possible. This required visiting the companies to contact occupational hygienists and other key staff. This study included exposure data on oil mist and oil vapour from all the companies currently involved in drilling on the Norwegian continental shelf.
Most of the reports compiled are from the past decade. The reasons for the few reports from the 1980s and early 1990s are probably less sampling activity, less focus from the authorities, fewer results available due to inaccessible data systems and loss of company history because of retirement or key personnel changing positions. Prior to 1991, no results were accessible from movable drilling rigs. The number of exposure measurements increased from 1989 to 2004, presumably reflecting increased monitoring activity with time. However, some reports, especially from the earliest years, were not expected to be accessible during the collection process. The reports have varying amounts of information, and few provide detailed data on the design of the mud-handling areas, the ventilation system, the physiochemical characteristics of the base oils used and the detailed work tasks. Thus, the models presented in this study are based on the rather coarse set of variables stated in most of the monitoring reports.
Sampling has traditionally been aimed at covering the expected worst-case conditions indicated by process parameters such as mud temperature and section of the well. Thus, the exposure data presented are probably higher than would be expected from representative sampling during the complete drilling of a well. On the other hand, one cannot exclude that the drilling conditions might have changed due to unexpected circumstances before the occupational hygienist arrived on the platform. This could have led to measurements during conditions deviating from the planned worst-case strategy. Further, the relevance of comparing 12 h limit values with individual 2 h samples is also questionable, and no attempt has been made to estimate total shift exposure.
In the mixed-effects models only the 2 h samples taken by the filter/charcoal tube method were included. Since 1989 the majority of measurements has been done by this method, which is considered to be the most accurate and efficient sampling method for oil mist and oil vapour. Full shift sampling methods for oil vapour reported to be used before 1989 (charcoal tubes or passive dosimeters), will probably be affected by the aerosol phase leading to an overestimation of vapour (Malvik and Børresen, 1988). The accuracy of the statistical exposure models is presumably improved by using only the results from the most standardized, industry specific methods for sampling and analysis. In addition, these models are expected to be more in compliance with the present exposure situations, and indicate effects of control measures taken during the last 15 years.
The fraction of oil mist might have been underestimated if hydrocarbons from the filter evaporated onto the subsequent charcoal tube. The sampling time of this method was limited to 2 h to reduce this factor. Studies on the evaporation of mineral oil metalworking fluids have shown losses up to 34% after 1 h of sampling for mineral oils with a viscosity of 4–5 mm2 s–1 at 40°C (Simpson et al., 2000). In an additional study Simpson (2003) concluded that material of light mineral oil samples can be lost when glass fibre filters are stored. No investigation has been carried out to quantify the analogous loss of mineral base oils used in oil-based drilling muds. Thus, the methods used in sampling and analysis should be further validated.
Although discussed in the literature (James et al., 2000; Gardner, 2003), the possible contribution of hydrocarbons from other sources like drilling mud additives or drilled cuttings has not been considered in this study. Furthermore, potential formation of hazardous substances like PAHs in the drilling mud caused by the effect of high pressure and temperature in the wells needs to be investigated.
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
This study aimed at describing the historical, personal exposure to airborne hydrocarbon contaminants in the form of oil mist and oil vapour in the mud-handling areas of offshore drilling facilities operating in Norwegian waters when drilling with oil-based muds. Although the exposure to air pollutants declined from 1979 to 2004, levels exceeding the Norwegian OELs are still measured. Thus, further control measures to reduce the emission of oil vapour and mist in the mud-handling areas need to be initiated in this industry.
Linear mixed-effects models were created to identify time trends and significant determinants of exposure between 1989 and 2004 when the glass fibre filter and charcoal tube sampling method was used. The models showed a declining time trend for both oil mist (6%) and oil vapour (8%). The type of rig, the mud temperature, technical control measures, type of base oil, viscosity of the base oil, work area and season of sampling appear to be associated with the exposure levels. Drilling crews on movable drilling rigs experience twice the concentrations of oil mist and oil vapour than workers on fixed drilling facilities. The levels of hydrocarbon air contaminants increase as mud temperature increases, and will reach high concentrations compared to Norwegian OELs, especially for oil mist. Research on these determinants of exposures is scarce, implying that further studies are needed on evaluation of technical control measures, characteristics of oil-based mud and process conditions.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
We are thankful for all support and feedback from occupational hygienists associated with the offshore industry: Knut Grove and Esther Sætvedt (Kostad Bedriftshelsetjeneste AS), Trond Schei [ConocoPhillips Norge (CPN)], Bjørg Eli Hollund (X-Lab), Inger Margrethe Haaland (Norsk Hydro ASA) and Jorunn Kirkeleit (University of Bergen). Reagan James (CPN) and Dagrunn Dirdal (M-I Norge AS) have given skilled advice regarding base oils. Stein Atle Lie and Trond Riise of the University of Bergen and Hans Kromhout of Utrecht University have given valuable comments on the statistical models. The Norwegian Oil Industry Association funded this study, and we are grateful for all the support from Carsten Bowitz and his colleagues.
Received April 20, 2005; in final form August 2, 2005
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Arnesen U (1985) Nordiske ekspertgruppen for grenseverdidokumentasjon 57. Oljetåke [The Nordic expert group for criteria documentation of health risks from chemicals 57. oil mist]. Arbete och Hälsa; 13.
Burstyn I, Kromhout H, Kauppinen T et al. (2000) Statistical modelling of the determinants of historical exposure to bitumen and polycyclic aromatic hydrocarbons among paving workers. Ann Occup Hyg; 44: 45–56.
Davidson RG, Evans MJ, Hamlin JW et al. (1988) Occupational hygiene aspects of the use of oil-based drilling fluids. Ann Occup Hyg; 32: 325–32.
Eide I. (1990) A review of exposure conditions and possible health effects associated with aerosol and vapour from low-aromatic oil-based drilling fluids. Ann Occup Hyg; 34: 149–57.
Gardner R. (2003) Overview and characteristics of some occupational exposures and health risks on offshore oil and gas installations. Ann Occup Hyg; 47: 201–10.
Hornung RW, Reed LD. (1990) Estimation of average concentration in the presence of nondetectable values. Appl Occup Environ Hyg; 5: 46–51.
HSE. (2000) HSE offshore technology report—OTO 1999 089. Drilling fluids composition and use within the UK offshore drilling industry. London: Health and Safety Executive.
Hudgins CM. (1991) Chemical usage in North Sea oil and gas production and exploration operations. Houston: Petrotech Consultants Inc. p. 622.24:665.6:66.
James RW, Schei T, Navestad P et al. (2000) Improving the working environment and drilling economics through better understanding of oil-based drilling fluid chemistry. SPE Drilling Completion; 15: 254–60.
Malvik B, Børresen E. (1988) STF21 A88028. Målemetoder for oljetåke/damp [Methods for measuring oil mist and vapour]. Trondheim: SINTEF Teknisk Kjemi.
Ministry of Petroleum and Energy. (2004) Facts 2004. The Norwegian petroleum sector. Oslo: Ministry of Petroleum and Energy.
NIOSH (1994) NIOSH Manual of Analytical Methods (NMAM®), 4th edn. Washington: DHHS (NIOSH). Available at http://www.cdc.gov/niosh/nmam/.
Norwegian Labour Inspection Authority. (2003) Administrative normer for forurensning i arbeidsatmosfære 2003 [Administrative standards for pollution in the working environment]. Oslo: Norwegian Labour Inspection Authority (Best. nr. 361).
Norwegian Oil Industry Association. (1996) Criteria for selection and approval of drilling fluids with respect to effects on human workers and marine ecological systems. Stavanger: Norwegian Oil Industry Association.
Norwegian Petroleum Directorate. (2004) Offshore Norway 2003. Stavanger: Norwegian Petroleum Directorate.
Norwegian Petroleum Directorate. (2005) The NPD's facts pages. Stavanger: Norwegian Petroleum Directorate. Available at http://www.npd.no/engelsk/cwi/pbl/en, accessed 6 April 2005.
Petroleum Safety Authority Norway. (2004) Regulations relating to conduct of activities in the petroleum activities (the activities regulations): section 34. Stavanger: Petroleum Safety Authority Norway.
Rappaport SM, Goldberg M, Susi P et al. (2003) Excessive exposure to silica in the US construction industry. Ann Occup Hyg; 47: 111–22.
Rønneberg A, Skyberg K. (1988) Mortality and incidence of cancer among oil exposed workers in a Norwegian cable manufacturing company. Part I. Exposure conditions 1920–79. Br J Ind Med; 45: 589–94.[Web of Science][Medline]
Rønneberg A, Andersen A, Skyberg K. (1988) Mortality and incidence of cancer among oil exposed workers in a Norwegian cable manufacturing company. Part 2. Mortality and cancer incidence 1953–84. Br J Ind Med; 45: 595–601.[Web of Science][Medline]
Simpson AT. (2003) Comparison of methods for the measurement of mist and vapour from light mineral oil-based metalworking fluids. Appl Occup Environ Hyg; 18: 865–76.[Medline]
Simpson AT, Groves JA, Unwin J et al. (2000) Mineral oil metal working fluids (MWFs) – development of practical criteria for mist sampling. Ann Occup Hyg; 44: 165–72.
Skyberg K, Rønneberg A, Kamøy JI et al. (1986) Pulmonary fibrosis in cable plant workers exposed to mist and vapor of petroleum distillates. Environ Res; 40: 261–73.[Medline]
Strand LÅ, Andersen A. (2001) Kartlegging av kreftrisiko og årsaksspesifikk dødelighet blant ansatte i norsk offshorevirksomhet. Innsamling av bakgrunnsdata og etablering av kohort [Survey of cancer risk and cause-specific mortality of Norwegian offshore oil industry workers. Collection of background data and establishment of a cohort]. Oslo: Cancer Registry of Norway.
van Tongeren MJA, Kromhout H, Gardiner K. (2000) Trends in levels of inhalable dust exposure, exceedance and overexposure in the European carbon black manufacturing industry. Ann Occup Hyg; 44: 271–80.
Vermeulen R, de Hartog J, Swuste P et al. (2000) Trends in exposure to inhalable particulate and dermal contamination in the rubber manufacturing industry: effectiveness of control measures implemented over a nine-year period. Ann Occup Hyg; 44: 343–54.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K. S. Creely, H. Cowie, M. Van Tongeren, H. Kromhout, J. Tickner, and J. W. Cherrie Trends in Inhalation Exposure A Review of the Data in the Published Scientific Literature Ann. Hyg., November 1, 2007; 51(8): 665 - 678. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Steinsvag, M. Bratveit, B. E Moen, and H. Kromhout Inter-rater agreement in the assessment of exposure to carcinogens in the offshore petroleum industry Occup. Environ. Med., September 1, 2007; 64(9): 582 - 588. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Steinsvag, M. Bratveit, and B. E Moen Exposure to carcinogens for defined job categories in Norway's offshore petroleum industry, 1970 to 2005 Occup. Environ. Med., April 1, 2007; 64(4): 250 - 258. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||