Data Loading...
A Characterization of Solution Gas Flaring in Alberta Flipbook PDF
Johnson, Kostiuk, and Spangelo 1168 Journal of the Air & Waste Management Association Volume 51 August 2001 emergenc
125 Views
42 Downloads
FLIP PDF 466.52KB
Journal of the Air & Waste Management Association
ISSN: 1047-3289 (Print) (Online) Journal homepage: http://www.tandfonline.com/loi/uawm20
A Characterization of Solution Gas Flaring in Alberta M.R. Johnson , L.W. Kostiuk & J.L. Spangelo To cite this article: M.R. Johnson , L.W. Kostiuk & J.L. Spangelo (2001) A Characterization of Solution Gas Flaring in Alberta, Journal of the Air & Waste Management Association, 51:8, 1167-1177, DOI: 10.1080/10473289.2001.10464348 To link to this article: http://dx.doi.org/10.1080/10473289.2001.10464348
Published online: 27 Dec 2011.
Submit your article to this journal
Article views: 317
View related articles
Citing articles: 8 View citing articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=uawm20 Download by: [50.116.19.84]
Date: 03 April 2016, At: 02:56
Johnson, andAssoc. Spangelo ISSN 1047-3289 J. Air Kostiuk, & Waste Manage. 51:1167-1177
TECHNICAL PAPER
Copyright 2001 Air & Waste Management Association
A Characterization of Solution Gas Flaring in Alberta M.R. Johnson and L.W. Kostiuk Combustion and Environment Group, Department of Mechanical Engineering, University of Alberta, Edmonton, Alberta, Canada
Downloaded by [50.116.19.84] at 02:56 03 April 2016
J.L. Spangelo Alberta Energy and Utilities Board, Calgary, Alberta, Canada
ABSTRACT Information reported here is the result of a detailed analysis of data on flared and vented solution gas in the Province of Alberta in 1999. A goal of characterizing these flares was to aid in the improved management of solution gas flaring. In total, 4499 oil and bitumen batteries reported flaring or venting with a combined gas volume of 1.42 billion m3. There was significant site-to-site variation in volumes of gas flared or vented, gas composition, and flare design. Approximately 5% of physical batteries generate 35.7% of the gas flared and vented from oil and bitumen batteries. Therefore, if one were to attempt to mitigate flaring, significant progress could be made by starting with only the largest sites. The monthly variability of gas volumes was considered because high variability could affect implementation of alternative technologies. It was found that slightly more than 40% of the sites were reasonably steady and had monthly deviations of 100% or less from the average flared volume. The variability in monthly volumes was less for the larger batteries. Data from individual well sites show significant variability in the relative concentrations of each of the major species contained in solution gas. IMPLICATIONS The Province of Alberta is one of North America’s largest oil-producing regions. The associated flaring and venting of more than 1.4 billion m3 of solution gas per year raises significant environmental issues for the general public and the energy industry. Currently, there is a lack of readily accessible information vital to characterizing the numbers, volumes, and composition of solution gas flares operating in Alberta. The data presented here will provide some of this missing information and will be of use to regulators, industry, researchers, and engineers in their quest to better manage solution gas flaring.
Volume 51 August 2001
INTRODUCTION Flaring is the process of disposing of unwanted flammable gases and vapors by combusting them in a flame in the open atmosphere. In a typical flare, air and fuel are not premixed and the combustion occurs as a turbulent diffusion flame in a crosswind. The purpose of a flare is to consume flammable gases and vapors in a safe, reliable, and efficient manner while converting them through oxidation to a more desirable emission than simply venting the gases to the atmosphere. Flaring is used extensively in the energy and petrochemical industries. Worldwide, it is estimated that 101.9 billion m3 of gas was flared or vented in 1997.1 Many different designs and strategies for flaring have been developed to meet the widely different purposes and operating conditions that industry requires. In the petroleum industry, flaring can be roughly categorized under one of three broad headings: emergency flaring, process flaring, and production flaring. Emergency flaring typically occurs at large facilities such as refineries and gas plants where the primary concern is for the safety of the plant personnel and protection of the plant infrastructure. When an emergency situation arises, such as a fire, compressor failure, valve rupture, and so on, large volumes of flammable gas may have to be disposed of in a matter of seconds. Under these conditions, flow rates of gas through a flare can be very high, and exit velocities may approach sonic speed. Process flaring may also occur at refineries, sour gas plants, and petrochemical plants. At these facilities, gases that leak past relief valves are shunted to a process flare for disposal. Unlike emergency flares, process flares burn almost continuously at relatively low flow rates. However, during startup and shutdown or during the evacuation or blowdown of process units, gas flow rates can be significantly higher. An excellent overview of process and
Journal of the Air & Waste Management Association 1167
Downloaded by [50.116.19.84] at 02:56 03 April 2016
Johnson, Kostiuk, and Spangelo emergency flaring is provided in Jones.2 Brzustowski3 discusses some of the technology applied to larger flares and reviews the various criteria used in the design of flares, specifically with respect to flame shape and length. Production flaring is a broad heading in itself and refers to all types of flaring that occur in various levels of the upstream petroleum industry during the production of oil and gas fields. Within this category, flares can vary significantly. During the initial development of a gas well, gas may be flared at very high flow rates for a period of a few days in what is known as well testing. Well-test flares may be of comparable size to the emergency flares previously described. However, unlike process and emergency flares, in most cases well-test flares do not have significant engineering provisions for smoke suppression or enhanced flame stability. Significant flaring can also occur during the initial development of an oil well when all associated gas may be flared for a period of time until the gas is “conserved”. Within this context, “conserved” means the gases are collected and processed later to sales-grade natural gas or used for fuel at the battery site. If gas volumes are uneconomical to conserve, all of the gas produced may continue to be flared for the life of the well. Typically, these continuous flares involve relatively low gas flow rates and subsequently low exit velocities compared with well-test or emergencytype flares. The primary contributor to continuous gas flaring in the upstream oil and gas industry is solution gas flaring, which is the main interest of the paper. Solution Gas Flares The term “solution gas” is used to refer to the collection of gases that come out of solution when conventional and heavy oil is extracted from high-pressure reservoir conditions and reduced to near atmospheric pressure. Once brought to the surface, the oil is separated from any associated water and solution gas at a facility known as an oil battery. At the battery site, the oil is temporarily stored before being processed further. The water is re-injected into the reservoir of origin and the gases may be flared, vented, or conserved. To provide information on the scale of flaring and venting in Alberta, some overall volumes are presented below. Solution gas volumes referred to in this paper are gas volumes produced from oil or bitumen batteries. The origins of these data are discussed later in the paper when a more detailed analysis is provided. In 1999, 94% of the 23.7 billion m3 of solution gas produced in Alberta was conserved, while the remainder was flared or vented. Figure 1 shows a breakdown of the total volumes of gas flared and vented in various sectors of the upstream oil and gas industry in Alberta. Specific
1168 Journal of the Air & Waste Management Association
amounts of flaring and venting in each sector are shown in Figure 2. In total, 2.01 billion m3 of gas were flared and vented in 1999. Of that total, 71% (1.42 billion m3) was flared and vented as solution gas at oil and bitumen batteries. Thus, solution gas flaring is the most significant contributor to flaring and venting in the upstream petroleum industry in Alberta. The sheer volume of solution gas being disposed of makes it a significant concern to industry, regulators, and the public. Why Is Solution Gas Flared? In Alberta, the choice of whether to conserve or flare the solution gas at any particular battery site is part of the flare management framework administered by the Alberta Energy and Utilities Board (AEUB).4 At any given site, there are several reasons why flammable gases and liquids may
Figure 1. Volumes of gas flared and vented in Alberta in various sectors of the upstream oil and gas industry in 1999.
Figure 2. Relative amounts of gas flared and vented in various sectors of the upstream oil and gas industry in Alberta in 1999.
Volume 51 August 2001
Downloaded by [50.116.19.84] at 02:56 03 April 2016
Johnson, Kostiuk, and Spangelo be flared or vented and not conserved. These include safety; high, low, or intermittent gas flow; low energy density of the gas (low heating value); presence of H2S or other contaminants in the gas; proximity to available infrastructure; and economics. A recent study by Holford and Hettiaratchi5 considered simple economics of six potential alternatives to solution gas flaring. These were low-pressure gas collection (clustering), electrical generation using gas turbines or reciprocating engines, electrical generation using “miniturbines,” cogeneration, re-injection of gas with produced water, oxidation (biological and physical), and collection and processing. Although potential economic reductions were identified at sites with large flare volumes, several significant barriers to implementation were also identified. Included among these barriers was the quality of available field data for making economic assessments. Successful implementation of any strategy to reduce flaring or to improve the performance of existing flares relies on accurate basic information with which a strategy can be developed. One of the goals of this paper is to better characterize solution gas flaring to aid in the improved management of solution gas flaring in Alberta. Environmental Issues of Flaring and Venting Environmental issues of gas flaring are generally described in terms of efficiency and emissions. The flare efficiency is a measure of the effectiveness of the combustion process in fully oxidizing the fuel. In a typical solution gas stream, two different efficiencies may be relevant: the carbon conversion efficiency (also known as the combustion efficiency), which measures the ability of the flare to fully convert all HCs to CO2; and the sulfur conversion efficiency, which measures a flare’s ability to convert H2S to SO2. When inefficiencies occur, unburned fuel, CO, and other products of incomplete combustion (e.g., soot, volatile organic compounds, etc.) are emitted into the atmosphere. In the case of venting, both the carbon- and sulfur-based efficiencies drop to zero since none of the fuel is being converted to CO2 or SO2. If the flare stream contains CH4, the unburned fuel represents an increase in greenhouse gas emissions, since the global warming potential of CH4 is 21 times greater than that of CO2 by mass6 (7.7 times greater by volume). If the flare gas contains H2S, any unburned fuel emissions are potentially toxic. In addition, any combustion device that emits products of incomplete combustion can raise health concerns for animals and people. DATA REPORTING Gas Volume Data In Alberta, operators of oil and gas batteries are required to complete production reports and submit them to the
Volume 51 August 2001
AEUB on a monthly basis. In addition to providing information on the amounts of oil, gas, and water received, produced, and delivered, battery operators are required to report the volumes of gas flared or vented. Although AEUB guidelines do not specify how this gas is to be measured, Interim Directive 94-17 states that volumes of gas less than 500 m3 are to be reported with 20% uncertainty and volumes greater than this are to be reported with 5% or better uncertainty. Smaller flaring and venting volumes are typically estimated by first measuring the gas/oil ratio in the crude oil stream and then inferring the amount of gas flared or vented from the measured amount of oil produced. Larger-volume sites may use orifice plates or other measurement devices. The data presented in this paper have been derived from these monthly production reports, which are compiled and stored by the AEUB. A relational database was created to analyze these data under various criteria. Although the AEUB has distributed parts of these data for public use since 1999, not all batteries are included in the publicly released data.8 Data that are connected in any way to wells categorized as experimental are held confidential and are not published. This omission is deemed necessary to protect the economically sensitive production data from experimental wells. In 1999, 29 of the 8249 oil and bitumen battery sites in the province were classified as experimental. However, flaring and venting from these sites totaled 0.167 billion m3, which represented 13% more than the 1.266 billion m3 of gas that was flared and vented at nonconfidential sites. It should be noted that the data presented in this paper are complete and include data from these confidential experimental batteries. (Confidentiality was maintained throughout data processing by not accessing any information on battery locations or operators.) Although most oil and bitumen well sites are physically tied into batteries that report production, flaring, and venting data to the AEUB, there are cases where collections of bitumen wells that are not physically connected report data as a single entity. A single report for a collection of physically separated sites is known as a “paper battery.” Although it is estimated that paper batteries make up 1.5% of the total number of battery reports submitted to the AEUB, they account for 23.6% of the solution gas flared and vented annually. The inclusion of paper batteries in the data is a significant complication that hinders data analysis. While it is generally assumed that wells in a paper battery are located within a small geographic region (typically within a few miles), this is not necessarily the case. The influence of paper batteries on the data will be discussed later in this paper.
Journal of the Air & Waste Management Association 1169
Downloaded by [50.116.19.84] at 02:56 03 April 2016
Johnson, Kostiuk, and Spangelo Gas Composition Data Battery operators are not currently required to report the composition of gas being flared or vented on an ongoing basis. Thus, there is no direct way to determine the composition of solution gas flared and vented in Alberta. The AEUB does keep a separate record of composition of solution gas measured at individual oil wells. Although operators are not required to measure composition at well sites, current guidelines state that if the measurements are made, the operators are obliged to submit the data to the AEUB. A data file of 5614 solution gas analyses was obtained from the AEUB for use in this paper. This number of analyses represents ~11% of the 51,976 active and abandoned wells associated with solution gas batteries in the province. Although these analyses are not a random sample of all of the wells in the province, they are currently the best available data to use in estimates of the composition of gases flared and vented in Alberta. An important compositional issue is the presence of H2S in the flare gas, which is a known toxic gas and a strong odorant. Since the well analysis data are not directly connected to the battery site data, it is necessary to consider an alternate approach to estimate the amount of H2S being flared at battery sites. As part of the application to create a battery code for the purpose of production reporting, operators are currently obligated to indicate whether the gas is “sweet” or “sour” (a qualitative indicator of the presence of H2S in the gas). However, no guidelines are presently in place that define at what maximum level of H2S the gas is to be deemed sour instead of sweet. Operators are also required to provide an estimate of the maximum H2S concentration in the solution gas, but this maximum is not necessarily indicative of operating conditions at the time of approval. Moreover, during the operating life of a battery, as wells are added and removed, the composition of the flared and vented gas would be expected to change. AEUB guidelines state that if there is more than 10 ppm H2S in the solution gas at a battery, sour gas warning signs must be posted. If the signs are posted indiscriminately at sweet sites, companies may be subject to an enforcement process. As part of the battery inspection process, the AEUB will note whether a battery is sweet or sour based on the posted signage and the inspector’s knowledge of the area, although this is obviously not a quantitative measure of gas composition. FLARING AND VENTING OF SOLUTION GAS IN ALBERTA In 1999, there were ~8249 active oil and bitumen battery sites scattered throughout Alberta that produced a total of 59.4 million m3 of oil and 23.7 billion m3 of solution
1170 Journal of the Air & Waste Management Association
gas. Although most solution gas produced in Alberta is “conserved”, ~6% of these gases were flared or vented at the battery sites. In 1999, 3715 of these battery sites reported volumes of gas flared totaling 0.938 billion m3, while 1346 reported venting of gas totaling 0.485 billion m3. In total, there were 4499 oil and bitumen batteries in Alberta that reported flaring or venting during 1999 with a combined gas volume of 1.42 billion m3. It is this number of flare sites and the volumes of gas being flared that make the process of solution gas flaring an environmental concern for the public and oil and gas producers. Characteristics of Solution Gas Flares It is extremely difficult to describe solution gas flares in terms of a common set of characteristics. Although operators of oil and bitumen batteries are required to report total volumes of gas flared at battery sites on a monthly basis, other data, such as the composition of gas being flared, flare diameter, type of ignition system, type of liquid separation system, and composition of liquids in the separation system, are not reported. Furthermore, anecdotal evidence suggests that most of these parameters can vary widely from site to site. Physical Characteristics. Although there is no single set of physical characteristics that describe or define solution gas flaring operations, certain common features are believed to exist in that the flare stack height for solution gas is on the order of 10 m high and a common stack size is ~10 cm in diameter. Other important parameters, such as the volume flow rate and velocity of the solution gas exiting from the flare stack (Vj), can vary widely. Moreover, gas flow rates to the flare may not be steady, depending on the operation of the oil wells feeding the battery and the downstream operations of facilities such as a gas plant that may shut down periodically. Variations in the mean wind speed (U∞) at the site also continually alter the flaring conditions. The ratio of U∞/Vj could easily vary from 0 to 25, which produces flames that are either upright or bent over horizontally, and can significantly affect the performance of the flare.9 Furthermore, there is considerable variation in the design of the flare tip in terms of wind shrouds and automated ignition systems (e.g., electric spark or continual pilot) that are mounted around the exit of the flare stack. Battery Size/Volume Distributions. Figure 3 shows a histogram of the number of oil and bitumen battery sites in Alberta sorted by their total volumes of gas flared or vented during 1999. In total, there were 4499 battery sites in Alberta that reported flaring or venting during that year. The logarithmic scale on the horizontal axis of the
Volume 51 August 2001
Downloaded by [50.116.19.84] at 02:56 03 April 2016
Johnson, Kostiuk, and Spangelo histogram highlights the large variability in flow rates among individual battery sites. The disparity between the median volume flared or vented (~60,300 m3/year) and the mean volume (316,200 m3/year) highlights the skewness of the volume distribution. Also shown in Figure 3 are the cumulative distributions of both the number of battery sites and the total volumes of gas released at all sites sorted by the amount flared or vented at the battery sites. These cumulative distributions give an indication of the estimated size of the flaring operations at individual battery sites and illustrate the proportion of gas flared at these sites. For example, referring to the dashed line in Figure 3, it is apparent that 95% of battery sites flare and vent less than 1.08 million m3/year. Assuming this amount of gas was consumed in a continuously operating flare, this would equate to an exit velocity of 4 m/sec on a 10-cm-diameter flare stack. Of course, most of these battery sites flare amounts much less than 1.08 million m3/year and would be expected to operate intermittently at higher and lower flow rates than their average flow rate. Thus, it is estimated that typical exit velocities for solution gas flares would be less than 6 m/sec. Although 95% of battery sites in Alberta flare or vent less than 1.08 million m3/year, the solid line in Figure 3 shows that these 95% of batteries only generate 43.6% of the total gas flared and vented at all battery sites. Alternatively stated, 5% (or 225) of the battery sites in the province account for 56.4% of the gas flared and vented at oil batteries in Alberta. This observation has significant implications for strategies to manage and mitigate solution
gas flaring. If solutions were implemented at the largest 5% of battery sites, this would affect more than 50% of the gas flared and vented in Alberta annually. Unfortunately, the practicalities of such an implementation are not as simple as they might seem, because the data contained in Figure 3 are biased by the inclusion of paper batteries. Although there were only 98 paper batteries in 1999 that reported flaring or venting, these 98 groupings accounted for 23.6% of the gas flared and vented in the province, or 336.4 million m3. Thus, many of these largest 5% of sites may not actually be physically connected as a single gas source. With this limitation in the data, it is difficult to address the issue of site-to-site variability in battery size or devise strategies for the management of these gases. Therefore, it is useful to remove the paper batteries from the data and study a reduced data set of batteries that are known to physically exist. Figure 4 shows the same type of information as Figure 3, but only for physical battery sites (i.e., no paper batteries). There were a total of 4401 physical battery sites that reported flaring or venting in 1999 with a total gas volume of 1.09 billion m3. The mean volume released at physical batteries was 246,900 m3/year, whereas the median volume was 58,700 m3/year. Thus, the significantly skewed distribution noted in Figure 3 is still apparent. In fact, the cumulative distributions show quite similar trends as those observed in Figure 3. From the dashed line in Figure 4, it is apparent that 95.2% of the physical sites flare and vent less than 1 million m3 of gas/year. These same sites only generate 53.9%
Figure 3. Distributions of gas volumes flared and vented at reported individual battery sites in Alberta in 1999.
Figure 4. Distributions of gas volumes flared and vented at physical battery sites in Alberta in 1999. “Paper batteries” have been separated out of the data.
Volume 51 August 2001
Journal of the Air & Waste Management Association 1171
Downloaded by [50.116.19.84] at 02:56 03 April 2016
Johnson, Kostiuk, and Spangelo of the gas flared and vented at physical batteries annually. Alternatively stated, the largest 5% of the physical batteries (221 sites) flare and vent 47% of the gas released at physical sites in the province, which represents 35.7% of the total solution gas flared and vented at all sites (including paper batteries) in the province. If one looks at the largest 20% of physical sites (881 sites), it is apparent that these 20% release 78.5% of the gas at physical sites, or 59.9% of all the solution gas flared and vented in Alberta. Thus, the significance of these distributions with and without paper batteries is the same—if one were to attempt to mitigate problems associated with flaring, significant progress could be made by starting with the largest sites in the province. Although the decision to concentrate on the largest sites might seem obvious at this point, other important complications have the potential to limit the applicability of alternatives to flares. One such problem is the variability of the volume of gas flared or vented at individual sites over time. Figure 5 contains an array of bar graphs that show the volumes of gas flared or vented on a monthly basis at four individual physical battery sites. Although these sites were chosen so that their total annual volumes were on the same order as the mean annual
volume for all physical sites of 247 × 103 m3/year, they show starkly different trends. Batteries can flare or vent fairly steady volumes (Figure 5a), have a single anomalous month (Figure 5b), demonstrate erratic monthly volumes (Figure 5c), or have erratic volumes with the occurrence of zero reported volume in some months (Figure 5d). On each plot in Figure 5, the total, average, and maximum gas volumes are noted. Also reported are the standard deviation of the monthly volumes and the deviation of the maximum, which is defined according to
(1)
One of the primary engineering advantages of flares is their ability to handle wide ranges of gas flow rates on a single device. In industry terms, this is often described as “turndown”, the ratio of the maximum sustainable flow rate to the minimum flow rate required for operation. Typically, the turndown of a flare is limited only by flame stability, and may be 100:1 or greater. Unfortunately, most other technologies, such as internal combustion engines,
Figure 5. Examples of month-to-month variation in volumes of gas flared and vented as reported at individual battery sites.
1172 Journal of the Air & Waste Management Association
Volume 51 August 2001
Downloaded by [50.116.19.84] at 02:56 03 April 2016
Johnson, Kostiuk, and Spangelo turbines, and compressors, have much narrower operating ranges. A typical turndown ratio for such a device might not be more than 2 or 3:1. In the context of flaring mitigation, this is a significant limitation. If a particular battery site has a relatively steady supply of gas, there may be many more possibilities to mitigate flaring than for a site with a highly variable gas supply. While turndown ratios required to handle all the monthly gas volumes at each site in Figure 5 could be presented, sites where the gas supply drops to zero in a given month (i.e., Figure 5d) would result in the turndown ratio going to infinity. Moreover, in a practical context, if one was to pursue an alternative technology to flaring at a given site, one might be willing to size the technology so that the maximum gas flow rates could be handled with the trade-off of venting or flaring gas when the flow rates were too low. With this in mind, a more useful parameter to describe the variability of the gas flow at a battery site is the deviation of the maximum monthly volume, where sites with smaller deviations are more suitable to implementation of any steady flow devices. The standard deviation is also a useful statistic, but it is more difficult to interpret with respect to a turndown ratio. The four plots in Figure 5 show different degrees of variability in the monthly volumes of gas flared or vented. Figure 5a is an example from a site where the flow rate of gas being flared appears to be essentially continuous. The standard deviation of the monthly volumes is only 4% and the deviation of the maximum is 9%. On the criteria of steady flow rate only, this site would be an example where alternate technologies might be applicable. By contrast, Figures 5b–5d all have some significant variability. The site in Figure 5b appears to have operated fairly steadily for 11 months of the year, but in June, 403% more gas was flared or vented. Upon closer examination of Figure 5b, one can determine that this battery normally conserves most of its gas. During the month of higher flaring, gas deliveries dropped, suggesting this may have been due to downstream facilities being shut down or for some other reason unable to handle the gas. In this example, 42% of the gas flaring or venting during the year occurred in a single month. It would be difficult to apply an alternate technology with limited turndown in this situation or at other sites with similar monthly volume distributions. Figures 5c and 5d show still different types of monthly distributions and appear to be almost random. The deviation of the maximum for the site in Figure 5c is 125%, compared with 210% from Figure 5d. The data in Figure 5d have the added feature that the site appears not to have operated (or at least did not flare or vent) for 3 months of the year. It is difficult to hypothesize which
Volume 51 August 2001
of many factors may be influencing the operation of the sites in Figures 5c and 5d. These four examples illustrate the wide variability that is apparent at solution gas battery sites. To characterize flaring and venting at battery sites in general terms, it is important to know what are the more prevalent types of monthly volume distributions; that is, what are the relative numbers of steady and unsteady sites. Figure 6 shows a histogram of the deviation of the maximum volumes for all 4401 physical battery sites. These results suggest that the deviation of the maximum monthly volume is less than 100% for more than 40% of the sites. At the same time, the tail of the distribution suggests that many sites have significant monthly variability. However, there is no way to tell from this figure if the monthly variability in gas volume is connected to either small- or largevolume sites. Figure 7 shows an array of plots where the deviation of the maximum data from Figure 6 has been separated (conditioned) by total annual volume to match the size bins (width of the bars on the histogram) from Figure 4. In this manner, it is possible to test the influence of size of battery on the monthly variability in volumes flared and vented. From Figure 7, it is apparent that the monthly variability is strongly connected to the size of the battery. It is clear that the smaller batteries have more monthly variability than the larger batteries. For example, less than 20% of batteries with annual flare and vent volumes between 1000 and 3162 m3 have monthly distributions such that the deviation of the maximum is less than 100%. However, as the range of battery size is increased in successive plots, the proportion of sites with deviations of
Figure 6. Distribution of the month-to-month variability in reported volumes of gas flared and vented at individual battery sites.
Journal of the Air & Waste Management Association 1173
Downloaded by [50.116.19.84] at 02:56 03 April 2016
Johnson, Kostiuk, and Spangelo
1 ≤ Tot. Vol. < 3.162
3.162 ≤ Tot. Vol. < 10
(x 103 m3)
(x 103 m3)
10 ≤ Tot. Vol. < 31.62
31.62 ≤ Tot. Vol. < 100
(x 103 m3)
(x 103 m3)
100 ≤ Tot. Vol. < 316.2
316.2 ≤ Tot. Vol. < 1000
(x 103 m3)
(x 103 m3)
1000 ≤ Tot. Vol. < 3162
3162 ≤ Tot. Vol. < 10000
3
3
(x 10 m )
(x 103 m3)
Figure 7. Variation in the month-to-month reported volumes of gas flared and vented at individual battery sites sorted by the size of the battery.
1174 Journal of the Air & Waste Management Association
Volume 51 August 2001
Johnson, Kostiuk, and Spangelo
Composition of Solution Gases. Apart from the physical and environmental differences among battery sites, there are significant variations in the composition and phase of materials being flared and vented. Unfortunately, as described in the section on Data Reporting, there is no prescribed reporting of composition analyses at individual battery sites. The AEUB does maintain a limited database of solution gas analyses from a selection of individual wells (i.e., basic gas phase composition of solution gas at individual oil wells), but these sites are generally removed from the battery sites, which often flare a blend of gases from more than one well. However, these analyses represent the best currently available data on solution gas composition and are used here to provide the best possible insight into the composition of gas being flared and vented at oil and bitumen battery sites in Alberta. Figure 8 shows a plot of the mean and maximum concentrations of major components found in solution gas at individual oil well sites. The most important observation to be made from these data is that there is significant variability in the relative concentrations of each of the major species contained in the solution gas. The notion
of an average composition of solution gas is essentially irrelevant. Even CH4 (C1 HCs), which dominates the average concentration of the gas analyses at nearly 70%, can vary greatly at individual wells. This variability is illustrated in Figure 9, which shows a histogram of C1 concentration for all 5614 solution gas analyses. Although the heavier HCs (C5, C6, C7+) are nearly negligible in the mean, at individual wells, their concentrations can be considerable. These variations in composition are significant in terms of the potential uses for the solution gas and of the ability of the gas to be combusted in a flare in a safe manner. Both the mass density (kg/m3) and the energy density (MJ/m3) of the flare stream are dramatically affected by the observed variations in composition. Assuming the composition of solution gas is alkane-based, the range of mass densities for the data in Figure 8 calculated at 273 K and 1 atm is between 0.65 and 2.89 kg/m3. The energy density (calculated as the higher heating value of the gas at 15 °C and 1 atm) varies significantly from 4.9 to 133.7 MJ/m3. Also, the presence of heavier HCs in the gases can dramatically affect the propensity of the gas to form soot when combusted. The inert compounds CO2 and N2, which have mean concentrations of about 4%, can make up significant portions of the solution gas, and at higher concentrations could be expected to reduce the energy density of the gas. Research on the efficiencies of gas flares has shown that energy density can have a dramatic effect on flare performance.9 Perhaps the most important component in distinguishing gas compositions at individual sites is the amount
Figure 8. Analyses of solution gas at individual oil well sites in Alberta.
Figure 9. Histogram of C1 HC concentration in solution gas at individual well sites in Alberta.
Downloaded by [50.116.19.84] at 02:56 03 April 2016
the maximum less than 100% steadily increases to nearly 70% for sites with annual volumes between 3.162 and 10 million m3. This result is significant since it suggests that larger battery sites are more likely to have steady volumes of gas flared or vented (at least as measured on a monthly basis). As discussed previously, these are the same sites that would need to be targeted first in any mitigation strategy for problems associated with flaring.
Volume 51 August 2001
Journal of the Air & Waste Management Association 1175
Downloaded by [50.116.19.84] at 02:56 03 April 2016
Johnson, Kostiuk, and Spangelo of H2S that may be present in the solution gas. Not only is H2S a toxic component in solution gas, it can also lead to severe corrosion of metal parts that come into contact with gas containing H2S or its combustion products. Figure 10 shows a histogram of H2S concentration in the available solution gas analyses. Because there is wide variability in the samples, the horizontal axes of Figure 10 are shown with a logarithmic scale. Although 1848 (32.9%) of the gas samples contained no H2S, its concentration is quite significant in some samples. In the context of flaring, gas that contains more than 10 ppm H2S is typically referred to as “sour”, while gas with less than this amount is referred to as “sweet”. Unfortunately, there is no reliable way to connect the available composition data with the flaring and venting volume data from individual battery sites. Thus, it is not currently possible to reliably estimate the amounts of H2S being flared and vented at solution gas batteries in Alberta. However, since this is an important issue to consider in a characterization of solution gas flaring, it is useful to make a qualitative estimate of the proportions of sweet and sour battery sites and gas volumes using AEUB site inspection data. Based on these data, it is estimated that 24% of the battery sites in the province flare or vent gas that is sour. By correlating these qualitative labels with the volumes of gas flared at individual sites, it is estimated that 36% of the gas flared and vented in the province is sour. Finally, it is generally accepted that some amount of liquid eludes the liquid knockout system and is carried to the flare in the form of liquid droplets entrained in the flare gases. However, field data on the composition, mass
Composition Analysis for 5614 Wells Number of Wells with Stated H2S Fraction
Figure 10. Histogram of H2S concentration in solution gas at individual well sites in Alberta.
1176 Journal of the Air & Waste Management Association
fraction, and size range of these liquid droplets do not exist. Given that solution gas is typically dissolved in an oil/water mixture in underground formations, it is probable that the entrained liquids are a mixture of brackish water and heavier HCs (C5–C20) that would be expected to vary from site to site. Since the presence of entrained liquid droplets in the flare gases has the potential to significantly affect flare performance, measurements need to be undertaken to assess the composition of these droplets in the field. CONCLUDING REMARKS In total, there were 4499 oil and bitumen batteries in Alberta that reported flaring or venting during 1999 with a combined gas volume of 1.42 billion m3. This volume of gas represents ~71% of all of the flaring and venting reported in the upstream oil and gas industry in Alberta. Despite the prevalence of solution gas flaring and venting and the implications of this practice for industry, regulators, and the public, the quality of currently available field data is poor. Data are complicated by the inclusion of “paper batteries,” collections of physically disconnected sites that are reported as a single “paper battery,” which represent 23% of the gas volumes flared and vented. Also, the composition of the gases being flared or vented at battery sites is currently not reported. From analyzing the best available data, it is clear that solution gas flaring is neither well defined by a single set of operating parameters nor characterized by any single battery site. There is significant site-to-site variation in volumes of gas flared or vented, gas composition, and flare design. Concentrating only on oil and bitumen batteries that physically exist (i.e., ignoring paper batteries), it is apparent that the distribution of volumes of gas flared and vented at individual batteries is highly skewed with a mean volume of 246,900 m3/year and a median volume of 58,700 m3/year. Using these same data, it is apparent that 95% of these sites flare and vent less than 1.0 million m3/year of solution gas. However, the remaining 5% of physical batteries (221 sites), some of which handle more than 5 million m3/year of gas, generate 35.7% of the gas flared and vented annually at all oil and bitumen batteries in Alberta. Similarly, the largest 20% of batteries (881 sites) flare and vent nearly 60% of the solution gas flared and vented in the province. If one were to attempt to mitigate problems associated with flaring, significant progress could be made by focusing on the largest sites in the province. An important characteristic of solution gas batteries is the monthly variability of the gas volume that is flared or vented. High variability in gas flow rate is a primary barrier for the successful implementation of alternative
Volume 51 August 2001
Downloaded by [50.116.19.84] at 02:56 03 April 2016
Johnson, Kostiuk, and Spangelo technologies to flaring. Although examples from individual sites show that some batteries have consistent monthly flare and vent volumes, others can deviate widely. Using the deviation of the maximum monthly volume from the average monthly volume, it was found that on average slightly more than 40% of sites could be considered “steady” (the deviation of the maximum from the average is less than 100%). However, the variability in monthly flow rate was shown to correlate with the size of the battery, and larger batteries (larger annual flare and vent volumes) were shown to have a higher proportion of “steady” sites. The current state of knowledge about the composition of gases being flared and vented in the province is much less complete. Using data from solution gas analyses at individual well sites as an indicator, it was shown that there is significant variability in the relative concentrations of each of the major species contained in solution gas. The notion of an average composition of solution gas is essentially irrelevant. Perhaps the most important component in distinguishing gas compositions at individual sites is the amount of H2S that may be present in the solution gas. Unfortunately, since the available well analyses are generally removed from the battery sites that often flare a blend of gases from more than one well, it is not possible to make reliable quantitative estimates of the amounts of H2S being flared and vented at solution gas batteries in the province. However, using AEUB site inspection data, it is qualitatively estimated that 24% of the battery sites in the province flare or vent gas that is sour (>10 ppm H2S). By correlating these qualitative labels with the volumes of gas flared at individual sites, it is estimated that 36% of the gas flared and vented in the province is sour. ACKNOWLEDGMENTS The authors acknowledge the support of Environment Canada, the Canadian Association of Petroleum Producers, and the Alberta Energy and Utilities Board.
Volume 51 August 2001
REFERENCES 1. 2. 3. 4. 5.
6.
7. 8. 9.
International Energy Annual 1998; DOE/EIA-0219(98); Energy Information Administration, Office of Energy Markets and End Use, U.S. Department of Energy: Washington, DC, 2000. Jones, H.R. Pollution Control in the Petroleum Industry; Noyes Data Corp.: Park Ridge, NJ, 1973; p 322. Brzustowski, T.A. Flaring in the Energy Industry; Prog. Energy Combust. Sci. 1976, 2, 129-141. GUIDE 60: Upstream Petroleum Industry Flaring Requirements, 1st ed.; Guide Series; Alberta Energy and Utilities Board: Calgary, Alberta, Canada, 1999; p 75. Holford, M.R.; Hettiaratchi, J.P. An Evaluation of Potential Technologies for Reducing Solution Gas Flaring in Alberta; Report to the Clean Air Strategic Alliance (CASA) Flaring Project Team, University of Calgary: Calgary, Alberta, Canada, 1998; p 31. Houghton, J.T.; Meira Filho, L.G.; Callander, B.A.; Harris, N.; Kattenberg, A.; Maskell, K. Climate Change 1995: The Science of Climate Change; Intergovernmental Panel on Climate Change (IPCC), Cambridge University Press: Cambridge, U.K., 1996; p 572. Interim Directive ID 94-1: Measurement of Oil, Gas, and Water Production; Alberta Energy and Utilities Board: Calgary, Alberta, Canada, January 18, 1994. Crude Oil and Crude Bitumen Batteries: Monthly Flaring, Venting and Production Data; ST-60 series data files; Alberta Energy and Utilities Board: Calgary, Alberta, Canada, 2000. Johnson, M.R.; Kostiuk, L.W. Efficiencies of Low-Momentum Jet Diffusion Flames in Crosswinds; Combust. Flame 2000, 123, 189-200.
About the Authors Matthew R. Johnson is currently completing his Ph.D. in mechanical engineering at the University of Alberta. His Ph.D. research is on the efficiencies of gas flares in crosswinds, and he is also the project engineer for the University of Alberta Flare Research Project, Combustion & Environment Research Group, 4-9 Mechanical Engineering Bldg., University of Alberta, Edmonton, Alberta, Canada T6G 2G8. Jim L. Spangelo is a senior engineer in the Operations Group of the AEUB, 640 5th Ave. SW, Calgary, Alberta, Canada T2P 3G4. He has participated in the development of Guide 60 Upstream Petroleum Industry Flaring Requirements and the AEUB’s annual flaring report. Larry W. Kostiuk is a professor of mechanical engineering at the University of Alberta. He is actively involved in combustion research and is the project coordinator of the University of Alberta Flare Research Project, Combustion & Environment Research Group, 4-9 Mechanical Engineering Bldg., University of Alberta, Edmonton, Alberta, Canada T6G 2G8.
Journal of the Air & Waste Management Association 1177