OCR Text |
Show American Flame Research Committees - AFRC 2018 Industrial Combustion Symposium University of Utah, Salt Lake City, Utah September 17-19, 2018 Real-Time Measurement of Industrial Gas Flare Emissions via UAS Technology Joseph D. Smith, Ph.D., Laufer Endowed Energy Chair Missouri University of Science and Technology, Rolla, MO Robert E. Jackson, Zachary P. Smith, Steve Rusakiewicz Elevated Analytics, Inc., Provo, UT Abstract Industrial Gas flares are used world-wide to reduce safety concerns in up-steam and downstream production of hydrocarbon products. These flares are generally classified as non-assisted utility flares, steam-assisted flares, air-assisted flares, pressure-assisted flares, enclosed flares, liquid flares and pit flares. These flares are designed to operate for very low hydrocarbon flow rates (due to fuel cost) and high hydrocarbon flow rates (due to safety constraints). They must also perform under variable ambient conditions (e.g., wind and rain) for non-uniform gas compositions. Hydrocarbon plants are often required to have a "Flare Minimization Plan" as part of their air permit. Plants that "routinely" flare gas also include flare gas recovery units to improve plant efficiency and reduce environmental impact. Flare stacks are designed to burn flammable gases high enough to minimize radiation flux to surrounding equipment and work areas and to reduce ground level concentrations of combustion emissions. Current technology used to monitor flare performance measure radiation levels, flare gas flow rates and compositions and ground level concentrations for CO, NOx, VOC's. Early work aimed at characterizing flare efficiency was limited to small single point elevated flares using extractive sampling techniques. More recent studies extended this by using techniques such as Differential Absorption LIDAR (DIAL), Open-Path Fourier Transform Infra-Red Spectroscopy (OPFTIR), passive FTIR (PFTIR) and Video Imaging Spectro-Radiometry (VISR). Data collected using these techniques were subject to temporally and spatially varying flare plumes from a single source. Results reported from this work typically are averaged which fails to capture the dynamic nature of flare operation under various ambient conditions. Also, none of these techniques have been applied to Multi-Point Ground Flares (MPGF) due to the size of the flare field and the associated sampling limitations. Elevated Analytics, Inc. has developed advanced sensor systems using various fast acting sensors to measure local gas concentrations, temperature and relative humidity of flare plumes. This data is reported from the UAS device(s) wirelessly to the ground monitoring station that can then be linked directly to the plant digital control system to effectively control flare operations as a function of plant operations. These sensor systems have been applied in various applications which will be discussed in this paper. Real-time spatially and temporally accurate data has been used to develop "time-varying" contour plots of local air quality and temperature. This feature allows the system to function as an early warning device as well and allows the plant to function at higher capacities without risking inefficient flare operation. Unmanned Aerial System Based Flare Emissions Monitoring AFRC 2017: Industrial Combustion Symposium 1. Introduction and Background Industrial Gas flares are used world-wide to reduce safety concerns in up-steam and downstream production of hydrocarbon products. As safety devices they allow a plant to vent flammable gases in the most environmentally friendly way possible. These flares are generally classified as non-assisted utility flares, steam-assisted flares, air-assisted flares, pressure-assisted flares, enclosed flares, liquid flares and pit flares. These flares are designed to operate for very low hydrocarbon flow rates to save on fuel costs but must also be able to operate at high hydrocarbon flow rates when needed for pressure relief venting of systems. The flames must also perform under variable ambient conditions including high wind and with high precipitation amounts. This must be accomplished with non-uniform vent gas compositions. As reported in previous work by the authors [1], regulatory requirements for flares in the United States (US) are contained in 40 CFR §60.18 and §63.11 and for other countries similar requirements are often used based on the US requirements. These requirements were developed from data gathered from a series of flare emissions tests led by the United Stated Environmental Protection Agency (US-EPA) from 1983 - 1986. [2], [3], [4] The goal of these US-EPA studies was to provide operating parameters for flares that would ensure efficient flaring with minimal emissions characterized by high combustion efficiency (CE). In the past, measuring the combustion products from a flare in order to determine CE was difficult and dangerous and recognizing these difficulties in being able to get actual emission measurements, regulations were centered around establishing operating parameters for likely to achieve high CE. Recent technological advances, however, have led to the development and application of remote sensing technology capable of measuring combustion products (i.e., carbon dioxide, carbon monoxide and select hydrocarbons) without the safety hazards related to physically sampling a flare plume. These new techniques use some form of optical spectroscopy to measure combustion products from a remote fixed station. Absorption LIDAR (DIAL), Open-Path Fourier Transform Infra-Red Spectroscopy (OPFTIR), passive FTIR (PFTIR) and Video Imaging Spectro-Radiometry (VISR) are techniques which have been tasted. Each of these systems includes a ground (or pole) mounted instrument which references the temporally and spatially varying flare flame and or the flare plume. Results reported from this work typically are averaged which fails to capture the dynamic nature of flare operation under various ambient conditions. Most of these techniques are affected by the atmospheric conditions and compensation for variations in these conditions is required. Most of these optical techniques are too expensive to be set up permanently for a single flare, but rather are specialty instruments that are used for spot checks at multiple sites and the associated setup, calibration, and testing time are typically expensive and disruptive to plant operation. Also, none of these techniques have been applied to Multi-Point Ground Flares (MPGF) due to the size of the flare field and the associated sampling limitations. These land-based optical techniques also present challenges for composite plumes from multi-burner flares, where conditions at different locations in the plume can vary. Page 2 of 10 Unmanned Aerial System Based Flare Emissions Monitoring AFRC 2017: Industrial Combustion Symposium A technique that can be set up quickly and could remotely measure flare emissions without significant interruption to plant operation remains the goal for both industry and environmental agencies. This paper describes a novel approach to use advanced sensor technology mounted onboard an Unmanned Aerial System (UAS) that can fly into the flare plume and remotely measure flare emissions as a function of position and time with this data then transmitted wirelessly back to the ground. The authors first presented this application at the AFRC conference in December 2017 [1]. This paper provides an update on the progress made in developing this new UAS system. 2. UAS Based Flare Emission Monitoring System The new UAS based flare emissions monitoring system described previously [1] has undergone continued development with numerous tests now completed. The system, EvA's EAGLE™ sensor system, has now been tested on several commercial flares and undergone two well documented tests of MPGF flare tips at Zeeco's flare test facilities in Broken Arrow, OK. Additional sensors have been added to the suite of possible sensors and new sensor partners have been added to the development team to provide more capability and flexibility. In addition to the carbon nanotube (CNT) sensors a novel new micro-electro-mechanical (MEMS) based sensor, referred as a Molecular Property Spectrometer (MPS), has been added which provides the capability of monitoring 12 different hydrocarbon gases all on a single sensor. This MPS sensor both identifies the hydrocarbon gas present and determines, at the percent of lower explosion limit (LEL) levels, the concentrations for the following gases: butane, ethane, ethylene, hexane, hydrogen, isopropanol, methane, pentane, propane, propylene, toluene and xylene, Also, a newly developed Tunable Diode Laser sensor is undergoing testing and showing great promise as a pipeline inspection platform when mounted on a fixed-wing UAS that offers significantly longer flight times that copter type UAS platforms. Current prototypes are based on existing commercially available UASs which have proven the concept but a temperature hardened UAS are also planned to extend the capability of the system to allow for measurements in hotter regions of the plume. These EvA EAGLE™ monitoring systems can be deployed rapidly in a very cost-effective manner to monitor any type of flare. This system can monitor even MPGFs well, as emissions in all areas of the plume can be determined. While a single EAGLE equipped UAS can be deployed and under quasi-steady state conditions can determine emissions throughout the plume, for larger plumes such as those from an MPGF, a swarm of multiple EAGLE-UAS devices can also be used to make multiple measurements at the same time as is illustrated in Error! Reference source not found.. E ach sUAS-CNT device transmits the sensor input to a ground station data acquisition system (DAQ). The rapid response time of the sensors combined with the possibility of monitoring at multiple locations at the same time allows for levels of detail in data from flares that have hitherto been unavailable. This level of detail will allow for: improved flare operation, improvements to CFD based flare models due to increased availability of data for validation, and improved flare designs. Page 3 of 10 Unmanned Aerial System Based Flare Emissions Monitoring AFRC 2017: Industrial Combustion Symposium Figure 1 - Multiple EAGLE sensor devices can be used to form a swarm to monitor emissions from an multiple points in MPGF plume simultaneously 3. Testing of the EAGLE system Testing has been done with the EAGLE system and results from three of the flare tests are presented below. 3.1. MPGF Burner Test, November 2017 Testing was done at Zeeco's flare test facility in Broken Arrow using a single MPGF burner tip firing natural gas. The test setup, as shown in Figure 2, had wind generally out of the southsouthwest at about 3 mph, but gusting from 1 to 4 mph with direction shifts of as much as 45 degrees quite common during the roughly 8 minute test duration. The test included both the UASEAGLE unit shown in Figure 3, and two fixed ground station EAGLE units placed downwind of the flare as shown in Figure 2. b. EAGLE fixed ground station a. Test setup Figure 2 - Test setup included the one UAS-EAGLE unit plus two fixed ground EAGLE sensor units placed downwind of the flare Page 4 of 10 Unmanned Aerial System Based Flare Emissions Monitoring AFRC 2017: Industrial Combustion Symposium Figure 3 - EAGLE system mounted on a DJI UAS platform The UAS-EAGLE system was launched prior to flare ignition. After ignition several radial passes using the flare as the center-point were made at various elevations to determine the size and position of the plume while data were being transmitted to the DAQ computer. The fluctuations in wind were very evident as the plume shifted its position. Data from the test are shown in Figure 5. The upper right image in Figure 5 indicates a plan-form view of two succeeding flight paths (arcs) flown by the UAS-EAGLE system. The temperature data (top left), CH4 data (bottom left) and CO data (bottom right) are shown as a function of location with the origin being placed at the flare Figure 4 -UAS-EAGLE system in location. The two flight paths indicated if Figure 5 were plume of flare. flown approximately separated by only about 60 seconds showed that the plume had shifted nearly 45 degrees between the to flight measurements. While, as may be anticipated for an efficient high-pressure MPGF burner, both CH4 and CO were very low (less than 10 ppm for both) with noticeable noise, there was a marked shift in the peak temperature location. This dynamic characteristic of the plume has made it inherently difficult to do measurements of flares. The agility of the UAS based EAGLE system makes it possible to track the plume. Automated flight with artificial intelligence (AI) plume tracking is under development which will enhance the UAS-EAGLE system. Data from the fixed-station EAGLE units is not shown as data from these two units showed no statistically significant departures from ambient air during the test. Page 5 of 10 Unmanned Aerial System Based Flare Emissions Monitoring AFRC 2017: Industrial Combustion Symposium Figure 5 - Data from the November 2017 test 3.2. Landfill Flare Test, January 2017 Testing was done at Montauk's landfill site in Sand Springs, OK. This site has a single utility flare that sees operation during times the Gensets are not in operation, usually due to scheduled maintenance operations. The vent gas is typical landfill gas composed of about 50% methane with high amounts of CO2. The site of the testing is shown in Figure 6 with the flare near the bottom of the photograph as shown. Near the flare (circled with a dashed line) is the drone, positioned downwind of the flare inside the plume. One of the exciting features of the EAGLE system is the special resolution of data from the plume. As shown in Figure 7 the data can be plotted in three dimensional space relative to the flare. The data points in Figure 7 are colored by temperature with the plume position well indicated by increased temperature in the downwind position. Page 6 of 10 Unmanned Aerial System Based Flare Emissions Monitoring AFRC 2017: Industrial Combustion Symposium Recipocating CAT engines burn Biogas from landfill to generate electric power Drone based sensor system Figure 6 - Landfill site for the January 2018 flare test Flare Figure 7 - Data points from EAGLE system plot in 3D Page 7 of 10 Unmanned Aerial System Based Flare Emissions Monitoring AFRC 2017: Industrial Combustion Symposium 3.3. MPGF Burner Test, June 2018 Testing was done at Zeeco's flare test facility in Broken Arrow using 8 MPGF burner tips firing propylene. This test marked the first use of the EAGLE system for a multiple burner test. The venting rate, including all 8 tips, was 34,615 lb/hr. This, plus the test setup and all other conditions for this test are shown in Figure 8. An image of the test with all 8 burners firing is provided in Figure 9. The UAS-EAGLE is visible in the upper portion of the image as it takes data in the flare plume. Figure 8 - Test conditions (left) and setup for the June MPGF burners test Figure 9 - MPGF burner test with EvA's UAS-EAGLE system, June 2018 Page 8 of 10 Unmanned Aerial System Based Flare Emissions Monitoring AFRC 2017: Industrial Combustion Symposium As with the previous flare testing, the fluctuating transient nature of the plume was very evident in the data gathered by the EAGLE system. A snapshot of the interactive data visualization tool is shown in Figure 10, where data is plotted relative to a topographical map (left) of the area around the flare. Each data point has all information gathered at that point in time and space and the user can access this by simply clicking on that point (see Figure 10.b). As can be seen, after a single arced flight path was able to locate the plume, the majority of the data were taken inside the plume as indicated by the grouped data points. As seen Figure 9, near the burner tips there were eight distinct flames but further up the flames tended to merge together. The wind was gusting from 5 to 12 mph during the test and visually it was evident that the flames broadened with more merging with the higher wind speeds. As is typical (see MPGF discussion above) the plume from all 8 burner tips was a merged complex plume with varying properties depending on location in the plume. a. Data plotted in 3D b. Example of expanded data from one point Figure 10 - Data from testing is provided in an interactive plotting package 4. Conclusions and Recommendations The new UAS based sensor system, EAGLE, has seen additional development with an expanded sensor suite now offering additional application. While sensing of flare emissions remains a core focus the EAGLE system shows great promise for other applications such as vagrant methane emissions from pipeline leaks, oil-fields, and landfills. Real-time flare plume sensing has been realized for industrial scale flares. Testing done with the EAGLE system has shown that the system works well and can track flare plumes even under fluctuating ambient conditions which result in a vary transient plume that can change drastically in location over a short period of minutes or less. The EAGLE system shows promise for monitoring emissions from all flares with high accuracy and selectivity and it can be used to monitor emissions from even large MPGFs which have hitherto been impractical to monitor. The UAS-EAGLE system's ability to track the fluctuating plumes holds great promise as this ability can be automated so that continuous pilot interaction will not be needed. Totally autonomous flights from takeoff to landing and including self-charging platforms, are currently under development which will make the EAGLE system another plant instrument Page 9 of 10 Unmanned Aerial System Based Flare Emissions Monitoring AFRC 2017: Industrial Combustion Symposium that requires very little human oversight and can monitor flare emission in real-time in an extremely cost-effective package. 5. Acknowledgments We acknowledge Elevated Analytics and especially George Singer for the generous support required to develop this game-changing technology that promises to improve our environment by monitoring air emissions from industrial facilities. We also acknowledge Zeeco for the use of their flare test facility to demonstrate this technology. 6. References [1] J. D. Smith, R. Jackson and Z. Smith, "Unmanned Aerial System Based Flare Emissions Monitoring," in AFRC 2017: Industrial Combustion Symposium, Houston, TX, 2017. [2] United States Environmental Protection Agency-Office of Air Quality Planning and Standards, "EVALUATION OF THE EFFICIENCY OF INDUSTRIAL FLARES: TEST RESULTS," EPA-600/2-84-095, May 1984. [3] United States Environmental Protection Agency, Office of Air Quality Planning and Standards, "EVALUATION OF THE EFFICIENCY OF INDUSTRIAL FLARES: FLARE HEAD DESIGN AND GAS COMPOSITION," EPA-600/2-85-106, September 1985. [4] United States Environmental Protection Agency, Office of Air Quality Planning and Standards, "EVALUATION OF THE EFFICIENCY OF INDUSTRIAL FLARES: H2S GAS MIXTURES AND PILOT ASSISTED FLARES," EPA-600/2-86-080, September 1986. Page 10 of 10 |