Refinery evaluation of optical imaging to locate fugitive emissions.

ABSTRACT

Fugitive emissions account for approximately 50% of total hydrocarbon emissions from process plants. Federal and state regulations aiming at controlling these emissions require refineries and petrochemical plants in the United States to implement a Leak Detection and Repair Program (LDAR). The current regulatory work practice, U.S. Environment Protection Agency Method 21, requires designated components to be monitored individually at regular intervals. The annual costs of these LDAR programs in a typical refinery can exceed US$1,000,000. Previous studies have shown that a majority of controllable fugitive emissions come from a very small fraction of components. The Smart LDAR program aims to find cost-effective methods to monitor and reduce emissions from these large leakers. Optical gas imaging has been identified as one such technology that can help achieve this objective. This paper discusses a refinery evaluation of an instrument based on backscatter absorption gas imaging technology. This portable camera allows an operator to scan components more quickly and image gas leaks in real time. During the evaluation, the instrument was able to identify leaking components that were the source of 97% of the total mass emissions from leaks detected. More than 27,000 components were monitored. This was achieved in far less time than it would have taken using Method 21. In addition, the instrument was able to find leaks from components that are not required to be monitored by the current LDAR regulations. The technology principles and the parameters that affect instrument performance are also discussed in the paper.

INTRODUCTION

"Fugitive emissions" is the term used to describe hydrocarbon leaks from valves, piping connections, pump and compressor seals, and other piping system components that occur as part of the normal wear and tear in plant operations. They are characterized as largely random occurrences of point-source releases into the atmosphere. Although the quantity of emissions is very small for a single component, the large number of components results in fugitive emissions being approximately half the plant total hydrocarbon emissions.

Since the late 1970s, regulations have existed to control these emissions. The Clean Air Act Amendments of 1990, as well as requirements in a large number of states, require all of the major sources of volatile organic carbon compounds and hazardous air pollutants to control fugitive emissions. As a result, nearly all of the refineries and petrochemical plants in the United States have implemented a Leak Detection and Repair (LDAR) Program to control fugitive emissions.

LDAR Requirements

Piping components (generally valves, pump seals, and compressor seals) are surveyed for fugitive emissions using a portable hydrocarbon leak detection instrument (organic or total vapor analyzer: OVA or TVA), according to U.S. Environmental Protection Agency (EPA) Reference Method 21. (2) Under EPA Method 21, the probe of a hydrocarbon leak detection instrument is placed at the leak surface of a component. Air and any leaked hydrocarbon are drawn into the probe and pass through a flame-ionization detector to measure the concentration of organic hydrocarbons. The instrument measures the hydrocarbon concentration in the airstream in parts per million by volume. If the measured concentration exceeds the regulatory definition of "leak" for the type of component being monitored, action must be taken to repair the leak within a certain number of days after it is identified. If the repair would substantially impact the operation of a process unit, it may be postponed until a scheduled maintenance shutdown. Records of each component and test must be maintained and made available for regulatory agency inspection.

Method 21 has several shortcomings. It does not measure actual mass emissions rates of the detected leaks. Empirical equations, used to convert parts per million by volume to mass rates, do not correlate well with actual measured emissions rates. In addition, Method 21 measures only the hydrocarbon gas that is drawn through its probe. It cannot distinguish between a point source leak and leaks from a crack or any combination of point sources and cracks. The routine monitoring that requires an operator to visit and screen each regulated component on a defined frequency and to identify the one component out of a thousand that is a large leak makes the process very labor intensive and inefficient. This effort can be quite expensive as well: in a typical U.S. refinery with more than 200,000 regulated components, the annual cost for an LDAR program can easily exceed US$1,000,000.

[FIGURE 1 OMITTED]

Smart LDAR

A study by the American Petroleum Institute (API) found that more than 90% of controllable fugitive emissions come from only approximately 0.13% of the components (3) and that these leaks are largely random. The results from that study are shown in Figure 1. The majority of the mass emissions come from a small number of components with high leak rates. A more efficient method for fugitive emissions monitoring would more cost-effectively find these large leakers. Optical leak imaging has been identified as an alternative to Method 21 to locate large leaks more efficiently. This technology was tested in a refinery environment after controlled laboratory calibration of the camera's detection limits. It has the potential to meet Smart LDAR principles, which are to more quickly scan components in a plant, identifying the large leakers, leading to lower survey costs and increased emission reductions.

OPTICAL LEAK IMAGING TECHNOLOGY

An emerging class of technology, generally referred to as optical leak imaging, offers an operator the ability to view leaking gas as a real-time video image. The remote sensing and instantaneous detection capabilities of optical imaging technologies allow an operator to scan areas of potentially leaking components much more quickly, eliminating the need to measure all of the components individually. Although many other technologies can detect the presence of hydrocarbons, optical leak imaging provides a real-time image of the gas plume and the equipment that allows identification of the exact source. (4)

There are two basic types of optical leak imaging technologies, based on "active" and "passive" detection. An active imager views a scene as it illuminates it with laser radiation having a wavelength that is absorbed by the gas to be detected. A gas becomes visible to the operator when it attenuates some of the backscattered laser light and appears in the image as a dark cloud. A passive gas imager views light that is passively radiated by a scene. This light can originate from reflected sunlight or from thermal radiation emitted by warm objects in the viewed area. A gas becomes visible in a passive image if its radiance differs from that of the background. This can occur if it absorbs backscattered sunlight or if it differs in temperature from the background objects. In both technologies, the leaking vapor appears as a cloud of "smoke" on a video display of the scene under inspection. The results presented here were obtained using an active imaging instrument.

Backscatter Absorption Gas Imaging

The principle of operation of the backscatter absorption gas imaging (BAGI) technology is the production of a video image of the area surveyed in the field of view produced by reflected (backscattered) laser light. (5) When the laser wavelength is strongly absorbed by the gas of interest, but weakly absorbed by atmospheric gases, the leak plume appears as a black cloud against the more brightly illuminated equipment background. Figure 2 shows a schematic of the BAGI technology principle. For aliphatic and olefinic hydrocarbons, BAGI illuminates the scene with infrared (IR) light while a "flying spot"-type IR camera detects the backscattered IR light in the field of view. The camera converts this backscattered IR light to an electronic signal, which is displayed instantaneously as a black and white image on the camera's viewfinder. Because the scanner is primarily sensitive to illumination from the IR laser source and not the sun, the camera is capable of displaying an image in either day or night conditions. The camera can be switched between visible and IR views, which allows the operator to differentiate between steam plumes and gas plumes (see discussion of atmospheric window, below).

Description of the Sandia National Laboratory Camera

The portable gas imaging device evaluated during this refinery study was a prototype developed by Sandia National Laboratory (SNL). The SNL camera views surroundings in IR light with the leaking gas appearing as a "black cloud" on a viewfinder or an attached video screen, as shown in Figure 3. The SNL camera is tunable in the 3.1-3.6 [micro]m wavelength range, where many hydrocarbons have strong absorption peaks.

[FIGURE 2 OMITTED]