|Dated: 10 Oct 2008
|Looking for Leaks
by Jeffrey H. Siegell , Ph.
January 8, 2008
Finding and repairing leaking components in process plants are essential to controlling fugitive emissions.
The current procedures used to locate leaking components in process plants are cumbersome and are not an effective use of resources. Leak inventory analyses show the potential to do better with a program focused on finding and repairing the largest leaks sooner. Optical imaging technology is available to quickly locate such leaks. Numerous successful applications of the technology have been completed at refineries and chemical plants, and techniques have been developed to quantify emissions and establish detection limits. In addition to being a replacement for traditional fugitives monitoring, optical imaging can be used to enhance plant safety and reliability.
Figure 1: Use of EPA Method 21 for monitoring piping components.
For many years, process plants have identified leaking components using a procedure called "sniffing." This procedure is based on using a sensitive gas sampling instrument to measure the concentration of hydrocarbons in the air adjacent to a potentially leaking piping component. It requires that an operator visit each component in a plant so that it can be individually monitored, as shown in Figure 1. This monitoring is very labor intensive and, since typically less than 1 or 2 percent of all components are found leaking, most of the effort appears to be wasted.
Specific details for properly applying EPA's test Method 21 are provided at www.tweekon.fr. Annual costs to conduct a leak detection and repair (LDAR) program at a large U.S. facility often exceed $1 million. Because of the complexity of this methodology, and alternative interpretations of the requirements, there have been a number of reports and publications providing detailed guidance on the correct application procedures. There continues to be disagreement, however, on the proper techniques to use for monitoring.
There are several limitations associated with the application of Method 21. These include the potential for false negatives (leaks missed) and false positives (unnecessary repairs) due to the significant variability of the ambient concentration for a specified mass leak rate.
This occasional inverse relationship of mass emission rate with Method 21 readings has been demonstrated during bagging studies in which components are enclosed to measure their mass emission rates. Figure 2 shows the potentially wide variation in readings for several mass emission rates based on actual field data.
Smarts, not sniffs
Enlarge this picture
Figure 2: Variability of Method 21 screening values.
In 1997, the American Petroleum Institute initiated work to identify a more effective LDAR methodology for refinery fugitive emissions control. The first step was to better understand the contemporary leak performance of typical refineries. Analyses from five years of monitoring at seven California refineries showed that over 90 percent of controllable emissions came from a little more than 0.1 percent of the piping components. This small group of components was the source of a vast majority of emissions.
Figure 3 shows the distribution of components and total emissions as a function of Method 21 screening values. Most components were found to be in the lowest screening range, where the contribution to mass emissions was negligible. Most of the emissions were from the very few components at the highest screening ranges. Over 92 percent of controllable emissions came from about 0.13 percent of the components.
Enlarge this picture
Figure 3: Distribution of leak rates and component counts.
The API Smart LDAR project had three key objectives. The first was to demonstrate a technology that could be used in a plant to cost-effectively find the high emission rate leaks. The second was to have federal regulatory requirements changed to allow the use of this new technology in lieu of Method 21 for locating leaking components. The third was to promote the commercialization of the technology to facilitate more widespread application in process plants for LDAR.
The institute's program included many individual projects aimed at developing and testing new technologies, demonstrating laboratory and field performance, and documenting emission control equivalency. This was a large cooperative effort involving several government agencies and the optical imaging industry as shown in Table 1 (see bottom of article).
Enlarge this picture
Figure 4: Optical imaging for detecting fugitive emissions.
Initially, there was no cost-effective technology found to locate leaks more rapidly than Method 21. However, instruments to accomplish this using optical imaging of hydrocarbon plumes were just starting to become available as the API study of leak distributions was completed. As illustrated in Figure 4, this technology allowed the leaks to be seen in real time as smoke on a video monitor. Plant equipment was also visible, allowing quick identification of the exact leak point. Hydrocarbon gas leaking from a valve as seen through the instrument is shown in Figure 5.
There are two basic types of optical imaging technologies: natural infrared imaging, generally referred to as "passive imaging" and laser-illuminated imaging, generally referred to as "active imaging." A passive gas image is produced by the reflection of sunlight in the infrared region from the equipment, with the gas cloud absorbing infrared light and thus appearing darker. Additionally, the relative difference between the gas cloud radiance and the background behind the gas cloud creates a contrasting image of the gas. With active imaging, the equipment is illuminated with infrared laser light, with the gas cloud image produced by the absorption of the laser light passing through and backscattered from the background behind the gas cloud.
In both technologies, the leaking vapor appears as a cloud of smoke on a video display of the scene being inspected. Both of these technologies have been commercialized and are in widespread use in process plants and the electrical transmission industry.
Some regulatory agencies in the U.S. are using optical imaging instruments in helicopters to search for leaks from processing plants and other industrial sources.
Emissions control equivalence
Figure 5: Leaking valve found with the optical imaging instrument.
To facilitate federal regulatory acceptance of Smart LDAR, the institute completed several analyses to demonstrate how to apply optical imaging and achieve equivalent or better emissions control than currently achieved with the existing regulatory requirements. This evaluation methodology was developed by the EPA and used Monte Carlo analysis to compare emission control method effectiveness. It required a demonstration that an alternative work practice (AWP) was equivalent in emissions control to that achieved by the current work practice (CWP).
The institute conducted several studies to develop sets of alternative monitoring frequencies and detection limits for an AWP equivalent to the currently required U.S. CWP. Its work evaluated current local, state and federal requirements for LDAR monitoring to identify combinations of equivalent imaging system detection sensitivities and monitoring frequencies. Leak rates were higher and the monitoring frequencies shorter than the current practices. For bimonthly monitoring (approximately every 60 days) a detection sensitivity of about 60 grams per hour was deemed to be equivalent.
Plant application requires a demonstration that leaks will be found with the optical imaging instrument. Rather than have each plant conduct comprehensive field testing, the institute combined an extensive laboratory wind tunnel test program with selective plant verification using bagging. This demonstrated detection limits for a variety of chemical species.
Several wind tunnel studies have been conducted to broaden the database of instrument detection limits. Initially, the tests were focused on olefins. These tests have shown that equipment performance is well within that required to show equivalent emissions control performance using an AWP. For example, while the required detection limit was 60 grams per hour for bimonthly monitoring, many of the demonstrated detection limits were in the 5 to 10 grams-per-hour range. In general, detection limits for aliphatic and olefin species tend to be lower than those for aromatic compounds.
Field studies have confirmed the applicability of the wind tunnel testing results. Leaks found in refineries and chemical plants with the optical imaging instrument were bagged to obtain actual mass emission rates. These fell well within the range of the wind tunnel tests demonstrating the usefulness of the wind tunnel tests to predict field performance. In several instances the relative inaccuracy of CWP was demonstrated when the actual mass emission rates from bagging were lower for some components with higher Method 21 readings.
Laboratory and plant testing have indicated that the optical imaging instrument is more sensitive for mixtures than would be predicted from the detection limits of the individual chemicals. This is important for refineries where most streams are complex mixtures of many chemical species.
A number of new technologies continue to be developed for optical imaging of hydrocarbon leaks. As the number of process plant applications multiply, these technologies will be commercialized and allow even wider application of Smart LDAR. Decreased emissions as well as improved safety and reliability will result. PE
References for this article are available upon written request to the editor.
Table 1: Major contributors to the development of a Smart LDAR
API Member Company Representatives (Technical review and project planning)
API Staff (Funding of technical studies and coordination of regulatory change activities)
Department of Energy - Office of Fossil Fuels (Funding development at Sandia Labs)
Department of Energy - Office of Industrial Technology (Funding development at Sandia Labs)
EPA - National Enforcement and Investigation Center (Participant in field testing)
EPA - Office of Air Quality Planning and Standards (Prepared regulatory change documentation)
Environ (Coordinated TCEQ chemical plant tests, participated in refinery tests)
Flir (Manufacturer of GasFind IR Camera)
Gas Imaging Systems, Inc. (Provided CO2 laser monitoring at field tests)
ICF Consulting (Conducted analyses of data and report preparation for EPA, TCEQ and API)
Laser Imaging Systems, Inc. (Manufacturer of GasVue camera and participant in field tests)
Leak Surveys Inc (Participant in field tests)
Pacific Advanced Technologies (Manufacturer of Sherlock camera and participant in field tests)
National Advisory Council for Environmental Science and Technology (EPA advisory group that funded early analyses - now dormant)
Sage Environmental (Performed bagging during field tests)
Sandia National Laboratories (Developed active IR camera and participated in field tests)
Texas Commission on Environmental Quality (Promoting development to find VOCs for ozone SIP)
Texas Council on Environmental Technologies (Funded additional testing of alternative imaging systems)
URS Radian (Performed bagging emissions quantification during filed tests)
Jeffrey H. Siegell , Ph.D.
Jeffrey H. Siegell, Ph.D., is a senior engineering associate with ExxonMobil Research and Engineering Co., Fairfax, Virginia, where he heads activities on estimating and controlling air toxic, hydrocarbon and odorous emissions. Dr. Siegell has been the coordinator of the API Smart LDAR project since its inception in 1997. He holds a BE(ChE) and a ME(ChE) from the City College of New York and a PhD(ChE) from the City University of New York.