Data from: Using controlled subsurface releases to investigate the effect of leak variation on above-ground natural gas detection
Data files
Nov 05, 2023 version files 152.15 MB
Abstract
Pipelines transport natural gas (NG) in all stages between production and the end user. The NG composition, pipeline depth, and pressure vary significantly between extraction and consumption. As methane (CH4), the primary component of NG is both explosive and a potent greenhouse gas, NG leaks from underground pipelines pose both a safety and environmental threat. Leaks are typically found when an observer detects a CH4 enhancement as they pass through the downwind above-ground NG plume. The likelihood of detecting a plume depends, in part, on the size of the plume, which is contingent on both environmental conditions and intrinsic characteristics of the leak. To investigate the effects of leak characteristics, this study uses controlled NG release experiments to observe how the above-ground plume width changes with changes in the gas composition of the NG, leak rate, and depth of the subsurface emission. Results show that plume width generally decreases when heavier hydrocarbons are present, the leak rate is reduced, and as leak depth decreases from 0.9 to 0.6 m. The above surface CH4 plume is undetectable when leaks are 1.8 m deep. As most survey methods typically prioritize leaks based on the leak size, this study shows that the effect of NG density on above-ground plume width is only 4%, equivalent to the effect of leak rate. This suggests that reported leaks in areas with heavier hydrocarbons could currently be missed or underestimated. Furthermore, this study shows that leaks from pipelines laid in covers meeting minimum depth requirements of 0.9 m could be easier to detect compared to those buried shallower. Overall, this study illustrates that leak survey protocols for flowlines and gathering lines should be different from distribution pipelines and tailored to the compositions of the transported NG to report emissions accurately.
README: Using controlled subsurface releases to investigate the effect of leak variation on above-ground natural gas detection
The dataset is a combination of methane measurements from the Ultraportable Greenhouse Gas Analyzer and meteorological data from the ultrasonic anemometer. This dataset is processed from the raw data in the following order:
- The meteorological data was converted from the epoch time to the George date; the original sonic data was split into separate dates.
- The 3D wind vectors u, v, and w were calculated from wind direction, wind speed, and wind elevation.
- The UGGA raw data were unzipped, converted to TXT files, then to CSV files.
- From the processed UGGA data, the following were extracted: Time,[CH4]_ppm, MIU valve, and MIU desc variables.
- For each experiment, multiple files were combined into daily files.
- The Monin-Obukhov length was calculated every 15 minutes and converted to Pasquill-Gifford classification.
Files
- Experiment_1.UGGA.MET.2022-04-25.csv: Day 1 data for the Baseline Experiment, 10 slpm of total natural gas (NG), 85% methane (CH4) 15% air, released at 0.9 m leak depth.
- Experiment_1.UGGA.MET.2022-04-26.csv: Day 2 data for the Baseline Experiment, 10 slpm of total NG, 85% CH4 15% air, released at 0.9 m leak depth.
- Experiment_2.UGGA.MET.2022-05-10.csv: Day 1 data for a gas composition experiment, 10 slpm of total NG, 70% CH4 30% ethane (C2H6), released at 0.9 m leak depth.
- Experiment_2.UGGA.MET.2022-05-11.csv: Day 2 data for a gas composition experiment, 10 slpm of total NG, 70% CH4 30% C2H6, released at 0.9 m leak depth.
- Experiment_3.UGGA.MET.2022-05-12.csv: Day 1 data for a gas composition experiment, 10 slpm of total NG, 70% CH4 10% C2H6 10% propane (C3H8) 10% butane (C4H10), released at 0.9 m leak depth.
- Experiment_3.UGGA.MET.2022-05-13.csv: Day 2 data for a gas composition experiment, 10 slpm of total NG, 70% CH4 10% C2H6 10% C3H8 10% C4H10, released at 0.9 m leak depth.
- Experiment_4.UGGA.MET.2022-04-04.csv: Day 1 data for a leak rate experiment, 5 slpm of total NG, 85% CH4 15% air, released at 0.9 m leak depth.
- Experiment_4.UGGA.MET.2022-04-05.csv: Day 2 data for a leak rate experiment, 5 slpm of total NG, 85% CH4 15% air, released at 0.9 m leak depth.
- Experiment_5.UGGA.MET.2022-06-02.csv: Day 1 data for a leak rate experiment, 1 slpm of total NG, 85% CH4 15% air, released at 0.9 m leak depth.
- Experiment_5.UGGA.MET.2022-06-03.csv: Day 2 data for a leak rate experiment, 1 slpm of total NG, 85% CH4 15% air, released at 0.9 m leak depth.
- Experiment_6.UGGA.MET.2022-05-25.csv: Day 1 data for a leak depth experiment, 10 slpm of total NG, 85% CH4 15% air, released at 0.6 m leak depth.
- Experiment_6.UGGA.MET.2022-05-26.csv: Day 2 data for a leak depth experiment, 10 slpm of total NG, 85% CH4 15% air, released at 0.6 m leak depth.
- Experiment_6.UGGA.MET.2022-05-27.csv: Day 3 data for a leak depth experiment, 10 slpm of total NG, 85% CH4 15% air, released at 0.6 m leak depth.
- Experiment_7.UGGA.MET.2022-04-12.csv: Day 1 data for a leak depth experiment, 10 slpm of total NG, 85% CH4 15% air, released at 1.8 m leak depth.
- Experiment_7.UGGA.MET.2022-04-13.csv: Day 2 data for a leak depth experiment, 10 slpm of total NG, 85% CH4 15% air, released at 1.8 m leak depth.
- Experiment_7.UGGA.MET.2022-04-14.csv: Day 3 data for a leak depth experiment, 10 slpm of total NG, 85% CH4 15% air, released at 1.8 m leak depth.
NAN Values
- The dataset contains NaN values for seconds when the analyzer did not report concentration values. This missing data is handled through listwise deletion. Due to the large dataset for the experiment, the listwise deletion does not significantly affect the overall result.
Variables
- Date: Date the experiment was conducted.
- Time: The time the drawn air was sampled.
- [CH4]_ppm: CH4 mole fraction in parts per million (ppm).
- dis_m: Distance from the release point to the mast in meters (m).
- wdirecenter_deg: Wind direction corresponding to the mast position from true North in degrees.
- hgt_m: Sampled height in meters (m).
- Air temp(degC): Air temperature in degrees centigrade.
- Air pressure(hPa): Air pressure in Hectopascal.
- RH%: Relative humidity.
- Sonic temp(degC): Sonic temperature in degrees centigrade.
- WD(degN): Wind direction from true North in degrees.
- WE(deg): Wind elevation in degrees.
- WS(m/s): Wind speed in meters per second.
- U(m/s): horizontal wind component calculated from the wind speed and wind direction in meters per second.
- V(m/s): horizontal wind component perpendicular to U calculated from the wind speed and wind direction in meters per second.
- W(m/s): vertical wind component calculated from the wind speed and wind elevation in meters per second.
- L: Monin-Obukhov length.
- PG: Atmospheric stability class <br>
Methods
The dataset is processed data for experiments that were conducted at conducted on the ‘rural testbed’ at Colorado State University’s Methane Emissions Technology Evaluation Center (METEC) in Fort Collins, CO, US, between April and June 2022. Four inlets were fixed at 0.5, 2, 5, and 7 m AGL on an 8 m stainless-steel mast 7 m northwest of the subsurface emission point. The four inlets were connected to the ABB LGR multiplexor via 30 m lengths of PTFE tubing (1/8" ID x 1/4" OD x 1/16" Wall Tygon® 2375 Ultra Chemical Resistant Tubing) which sampled each height for one minute sequentially. The outlet of the multiplexer was connected to a VACUUBRAND GMBH + CO KG MD1 vacuum pump and then tee-ed to an ABB LGR-ICOS GLA 132 Ultraportable Greenhouse Gas Analyzer (UGGA). The UGGA is a laser absorption spectrometer measuring methane, carbon dioxide, and water mole fractions in an air sample. It reports mole fractions each second, with a stated precision of 1.4 ppb at 1 Hz, 0 to 100 ppm linear measurement range, and 0.01 – 1 Hz measurement rate. Micrometeorological data is raw data measured at 10 Hz using RM Young 81000 ultrasonic anemometer installed 6 m above the ground at METEC.
The sampled methane mixing ratios from the UGGA were aggregated into micrometeorological data. The Monin-Obukhov length (L) was calculated from the surface friction velocity (u*, m s-1), the mean absolute air temperature (T, K), the von Kármán's constant (kv=0.41), the gravitational acceleration (g = 9.8 m s-2) and the 3D horizontal/vertical wind vectors (u, v, and w, m s-1), respectively. For analysis, the L was converted to Pasquill-Gifford stability class (PGSC), where PGSC A is extremely unstable (−100≤?<0), PGSC B/C is unstable (−500≤?<−100), PGSC D is neutral (|?|>500), PGSC E/F is stable (500≤?<100), and PGSC G is extremely stable (0<L≤100).