Steve Karcher, Ball Aerospace & Technologies
Developing a remote sensing instrument for real-time detection and quantification of fugitive natural gas emissions that must also adapt to evolving customer requirements driven by emerging industry regulations.
Using the timing and synchronization capabilities of the NI PXI platform, the integrated high-throughput I/O of a FlexRIO digitizer, and a LabVIEW-programmable FPGA to create the signal processing embedded system in a sophisticated differential absorption lidar product.
Steve Karcher - Ball Aerospace & Technologies
Phil Lyman - Ball Aerospace & Technologies
Jarett Bartholomew - Ball Aerospace & Technologies
The significant growth in the production, usage, and commercialization of natural gas is placing unprecedented demands on the nation’s pipeline system. The Pipeline and Hazardous Materials Safety Administration (PHMSA) develops and enforces regulations for the safe operation of the nation’s 2.6 million mile pipeline transportation system (U.S. Department of Transportation, 2016). Through PHMSA programs, serious pipeline incidents have decreased by 39 percent since 2009, according to the Department of Transportation (DoT). Recent incidents such as the 2010 San Bruno, California pipeline explosion and the 2015 Aliso Canyon gas leak are only two of more than 250 serious pipeline incidents since 2009.
Natural gas consists primarily of methane. Methane is the second most prevalent greenhouse gas emitted in the United States and accounted for about 11 percent of all US greenhouse gas emissions from human activities in 2014. Methane is emitted naturally and by human activities such as leakage from natural gas systems. The US Environmental Protection Agency says that the comparative impact of methane on climate change is more than 25 times greater than that of carbon dioxide over a 100-year period. Continued natural gas pipeline incidents and leaks, the associated impacts, and oil and gas industry regulations drive the need to promptly detect, classify, and resolve fugitive methane emissions.
Under funding from the PHMSA and Ball Aerospace, Ball used more than 50 years of remote sensing expertise to develop a system called Methane Monitor. Methane Monitor identifies methane emissions on the ground from a fixed-wing aircraft. Unlike existing methods of aerial leak survey, Methane Monitor operates from a single-engine, fixed-wing aircraft for lower cost than sensors mounted on helicopters. It images the full plume of methane gas as a more precise method of monitoring leaks, it can notify facility operators immediately of large emission sources, and it provides full reports within hours of the end of the flight. Development of these advantages placed large demands on high-throughput signal acquisition, synchronization, and processing.
In light detection and ranging (lidar) systems, a laser source emits a pulse of light. The pulse interacts with targets such as the ground or structures. Some of these interactions result in backscattered photons, which are collected and recorded as a function of time. This time-of-flight data directly corresponds with the range at which the scattering occurred, allowing generation of a 3D model of the illuminated topology.
Lidar range measurements are inherently part of differential absorption lidar (DIAL) measurements. DIAL operates at two laser wavelengths: one on-resonance and one off-resonance of a molecule of interest. Since the on-resonance wavelength is more strongly absorbed by the molecule, the difference between the two signals correlates to the amount of the molecule in the laser’s path. Thus, DIAL systems can measure the range and quantity of target molecules in the atmosphere (U.S. Department of Commerce, 2016).
DIAL systems look at sharp absorption lines in the spectrum, and Methane Monitor targets the methane molecule (CH4). We designed Methane Monitor so we could compare the resonance features uniquely from other molecules that might confuse the measurement. These measurements require a signal-to-noise ratio approximately 500 times better than what’s needed to establish range alone.
The environment imposes challenges because return signals are subject to changes in ground reflectivity. Imperfections in the laser impose challenges because the pulse energy and wavelengths of the two pulses vary independently across firings. Hence, Methane Monitor calibrates every measurement for background reflectivity and normalizes the received energy to the transmitted energy.
Methane Monitor also measures a calibrated methane sample before each target measurement. We can use the calibrated methane measurement to correct shot-to-shot instabilities in laser wavelength by reverse calculating the absorption constant.
Methane Monitor performs the background, reference, and receive measurements each time the laser fires. The on-resonance and off-resonance pulses are separated non-deterministically by a few hundred nanoseconds. The range depends on the customer’s survey objectives and aircraft’s attitude, and is generally 500 m to 1 km above ground level (AGL).
One or both return pulses may be well-formed (nominal), misshaped (for example, tall foliage), or absent (for example, massive concentration). When nominal return pulses exist, the lidar range measurement is combined with the DIAL measurement to provide a path integrated methane measurement. Depending on survey objectives, between 1,000 and 10,000 measurements are made per second.
Proper location and tracking of targets, attitude correction, and geo-location of methane measurements require synchronization between every component in Methane Monitor. We found that a PXI Express chassis provides the ideal platform for several key components.
Timing and Synchronization
Methane Monitor’s timing and synchronization centers on the PXI-6683H module, which includes a GPS-aligned system reference clock to the laser and PXI embedded systems. The system reference clock is available to all PXI Express peripherals.
The PXIe-6341 X Series DAQ uses reference clock synchronization to synchronize analog commands and telemetry.
The NI-5761 FlexRIO adapter module analog-to-digital (A/D) converters align to the system reference clock through IoModSyncClk and the sample clock select/commit signals in the socketed CLIP.
A FlexRIO FPGA module runs the custom digitizer and DIAL algorithms. The FPGA block diagram is synchronized to the system reference clock out of the box.
The PXI-6683H also generates asynchronous counter reset signals for the FPGA through PXI trigger lines. Counter values are packaged with each measurement. They can verify, geo-locate, and interpolate the measurements against data obtained from a position and orientation system (POS) and steering mirror controller.
The POS receives event triggers from the PXI-6683H through an SCB-68A device cabled to the PXIe-6341. The PXI-6683H also triggers the GigE context camera, which is otherwise automated using NI-IMAQdx.
The PXI-6683H generates a future time event that starts Methane Monitor shortly after liftoff.
The hardware, NI-Sync, NI-DAQmx, NI-IMAQdx, FlexRIO, and FPGA features make it possible for Methane Monitor to guarantee tightly synchronized and geo-located range and methane measurements across real-world operating conditions.
DIAL Signal Processing
The FlexRIO FPGA module runs one of several custom bit files depending on survey objectives. The FPGA captures the DIAL signals, measures range, and calculates methane concentration.
Data Serialization and Correction
Methane Monitor uses three or four transducers with the NI 5761 FlexRIO digitizer adapter module. The resulting throughput (between 11.2 Gibit/s and 15.2 Gibit/s) dictates FPGA processing. After calibrating each A/D converter count for the transducer chain, the FPGA performs the variable offset correction and serializes the high-throughput data by taking advantage of time between laser firings.
Custom Triggering
Pulses from each serialized signal are precisely acquired about the peak A/D converter count using level-triggered circular buffers. The serialization, custom triggering, and custom acquisition reduce the data throughput. Timestamps are assigned to each peak for the lidar range measurement.
DIAL Analysis
The FPGA performs several quality checks on the data. For example, it verifies that ground pulses were received, and it sets various flags based on pulse parameters. The FPGA reshapes each pulse to correct deterministic electrical effects. It executes Methane Monitor’s methane concentration algorithm every time the laser fires and streams telemetry to a LabVIEW application running on a PXIe-8135 controller.
The LabVIEW application provides the operator with an instantaneous view of the captured pulses, measurements, performance, system health, and more. The LabVIEW application serves the final data product to Ball Aerospace’s lidar visualization software that overlays the range and concentration measurement on the context camera image. All data is logged to an NI 8260 1.2 TB PXI SSD. We used DIAdem software to postprocess Methane Monitor’s data for quality assurance and continuous improvement.
Support From NI
Ball Aerospace has a heritage in space-based DIAL satellites such as CALIPSO, and airborne tactical lidars such as Total Sight. In contrast to these systems, a small team rapidly designed, implemented, tested, and evolved Methane Monitor. PXI electronics have supported several DIAL iterations and various survey objectives, including one for detecting liquid propane, and they continue to evolve for new survey objectives.
We engaged NI’s technical and field representatives regularly while specifying the PXI components. We used all aspects of NI’s high-throughput FPGA and FlexRIO customer education courses in Methane Monitor’s FPGA. The carefully designed synergies of NI’s PXI ecosystem facilitate Methane Monitor’s differentiating features at the lowest R&D cost imaginable.
Over 100 hours of flight time have been logged, and the methane detection threshold has been determined as a function of wind speed.
We have detected methane flow rates as low as 50 standard cubic feet/hour (SCFH). We can configure Methane Monitor’s sensing swath width up to 200 meters wide. The system has a spatial resolution and geo-location accuracy of better than 2 meters each.
Methane measurements are color-coded and superimposed on co-boresighted context images to provide a real-time view of methane emissions to the operator.
Data is prioritized after the flight, and additional postprocessing produces overlays compatible with Google Earth.
Ball plans to incorporate NI’s PXIe-5172 programmable digitizer in the next generation of Methane Monitor. That instrument can operate from ~3,000 ft AGL (two times higher altitude). It can survey a five times broader swath and offer two times better spatial resolution with comparable chemical sensitivity. The next-generation instrument can make cost-effective survey of highly branched, low pressure gas distribution assets possible. It can also facilitate leak survey of oil and gas gathering lines, which are broadly distributed across oil and gas formations.
The authors acknowledge the US Department of Transportation, Pipeline and Hazardous Materials Safety Administration for its support of DIAL instrumentation development for both hazardous liquids and natural gas: Contracts DTPH5613RA000001 and DTPH5615RA000016, respectively. Further funding was provided by Ball’s Commercial Aerospace business unit.
The authors further acknowledge the active guidance and support from pipeline operators that took an early interest in this advanced technology: Pacific Gas and Electric (PG&E) and the Colonial Pipeline Company.
Steve Karcher
Ball Aerospace & Technologies
Boulder, CO 80301