Transmission, Distribution, & Refining

Technologies to Enhance Operation of the Existing Natural Gas Compression Infrastructure

DE-FC26-02NT41646

The goal is to improve operation of integral engine/compressors by reducing fuel consumption, increasing capacity, and enhancing mechanical integrity, leading to increased reliability of the nation's natural gas transmission infrastructure.

Collage of 4 photos from Duke Energy Station showing Clark HBA Units

HBA Units at Duke Energy Station

Background:
This three-year project employed field testing, data acquisition, and data analysis to develop and evaluate a number of methodologies for balancing, operating and controlling the integral reciprocating engines/compressors used in pipeline gas transmission service. The primary approach was to test technologies that could balance the engine power cylinders and distribute load in the compressor cylinders in order to minimize fuel consumption, minimize damage rate, and maximize capacity. Alternative approaches such as peak-firing pressure balancing and standard deviation balancing were evaluated. Combinations of speed and compressor loading that minimize bending and torsional stress were evaluated using a model verified by crank strain measurements. To further control fuel consumption, the influence of compressor load distribution on trapped air fuel ratio through torsional vibration and its effect on port opening time was quantified and optimized for fuel flow.

The field testing was designed to acquire uniquely comprehensive data sets using customized equipment and methodologies. The data acquired included crankshaft bending strain, cylinder pressure on the engine and compressor, fuel flow, instantaneous crankshaft rotational velocity, bearing centerline vibration and dynamic pressures in manifolds. Tests were carried out at six different gas transmission compressor station sites. A field data system was designed to concurrently acquire multiple channels of data, including dynamically variable pressure in both power and compression cylinders. The data was analyzed and hypotheses developed on methods to optimize operation and performance of integral engine/compressor units. These hypotheses were then tested during subsequent repeat visits to the test sites.

Performers:
Southwest Research Institute (SwRI) – project management and research products

Location:
San Antonio, TX 78238

Potential Impact:
Through enhanced performance of the nation’s compression infrastructure significant impact could be achieved in several important areas. The increase in both compressor and engine efficiencies could bring about significantly increased gas throughput capacity for the existing gas infrastructure, reducing the need for new construction of facilities. In addition, operational enhancements minimize environmentally harmful emissions from the systems and decrease maintenance requirements and associated downtime, which helps increase the reliability of gas delivery through the pipeline network. Finally, enhanced performance and efficiency of the systems can reduce operating cost through reduced fuel and maintenance costs, which benefits end use consumers.

Accomplishments:

  • Carried out the first four of six field tests on reciprocating engine/compressor units using a suite of tools and methodologies designed to gather unique and comprehensive data sets. Engines tested have included two GMw-10s and two HBA-6Ts, and two stroke integral compressors,
  • Analyzed the resulting data and developed a number of insights that could lead to improvements in operational efficiency.

Observations from the suite of tests performed at that field sites include the following findings:

  • Fuel supply, trapped air, and their ratio vary from cylinder to cylinder. The industry widely uses fuel adjustment to balance combustion, but high cycle-to-cycle variation complicates this practice and limits the benefits of balancing. The most common method equalizes peak-firing pressures (PFP), but with unequal trapped air, unequal fuel/air ratios can result.
  • A newly invented method (compression pressure ratio (CPR) balancing) equalizes combustion pressure ratio (CPR is equal to the ratio of PFP to compression pressure) across cylinders and has proven feasible, with some evidence of reduced heat rate. Implementation involves calculating CPR each cycle, then averaging over multiple cycles. Cutting fuel to high CPR cylinders flattens the CPR distribution. PFP balancing works against compression pressure variation (observed at 6 percent -to 12 percent across cylinders), while CPR balancing works with this variation.
  • An alternative, with a similar goal, equalizes each cylinder's cycle-to-cycle standard deviation in PFP. Limited testing indicates this benefits crankshaft integrity.
  • Data shows 25 percent to 50 percent dynamic variability in manifold pressures, which likely contribute to air imbalance. Separate, ongoing tasks will characterize manifold dynamics and air imbalance in a GMVH6, and optimize manifold design to reduce this imbalance.
  • Global equivalence ratio can vary, and the project has shown feasibility of a low-cost means to maintain an equivalence ratio set point via turbocharger wastegate control.
  • Heat rate depends strongly on load, emphasizing the need for accurate, reliable, brake power measurement. Inaccurate torque can overload engines or hurt their efficiency. Inferential methods show up to 10 percent discrepancy; increasing their accuracy requires extensive mapping and errors can still result when malfunctions occur. The GMRC Rod Load Monitor (RLM) will measure power directly, optimize heat rate, and avoid overload.
  • Testing the evolving RLM has guided refinements; the project has now demonstrated the first self-powered RLM with digital telemetry on a large integral engine. Indicated power provides a viable calibration basis and initial evaluation showed the result remained consistent with indicated power when speed and load varied. The RLM measures torque upstream of ring/rider band losses, and translating RLM power into engine brake torque for set point comparison will demand better knowledge of such mechanical losses.
  • The crankshaft Strain Data Capture Module (SDCM) revealed: 1) how standard deviation balancing reduced the number of high strain excursions, 2) how reducing speed and high-pressure fuel cut strain, and 3) how advanced timing increased strain.
  • Because of sensitivity to small, natural, load variations, using heat rate directly to compare operational changes incurs uncertainty. However, the heat rate versus load chart shows promise as basis for comparison. While not conclusive, tests showed a distinct reduction in heat rate after CPR balancing (~100 BTU/HP-hr.), when compared to the baseline heat rate/load chart.
  • Similar comparison showed distinct heat rate reduction with two degrees HBA-6T timing advance. GMW10 timing tests showed the same or more heat rate reduction. Operators tend to limit timing advance for reduced heat rate because of the potential for detonation. A detonation detector loaned by Metrix shows promise as a sensitive and discriminating device to avoid detonation when advancing timing.
  • Comparing two different GMW10's showed much leaner operation, reduced heat rate, and reduced NOX concentration for a unit modified with high-pressure fuel and turbocharger. This unit also showed lower crankshaft strains and lower peak pressure.
  • Data shows system thermal efficiencies from 26.5 percent to 30 percent with “as found” timing. System thermal efficiency helps assess any complete compressor package. Attempts to assess how speed influences system efficiency emphasize the need to enhance mechanical efficiency knowledge and (ideally) avoid depending on assumed mechanical efficiency. Future rod load tests should help generate this knowledge.
  • Observed compressor thermal efficiencies (84 percent to 91 percent) have significance because high values reduce fuel consumption and increase capacity for fixed engine power.
  • Data confirms the value of monitoring discharge temperature to catch deteriorating compressor performance, and the potential role of other integrity monitoring methods using vibration and torsional velocity.

Observations from the initial suite of tests performed at the final two field sites focused on compressor operation and include the following findings:

  • Tests were performed on Duke’s Bedford Station and Dominion’s Groveport Station. Units tested were an HBA-6 with four compressor cylinders at Bedford (1,320 nominal HP; 300 RPM nominal speed) and a TCVC10 with three compressor cylinders at Groveport (5,000 nominal HP; 330 RPM nominal speed). Both candidates have the potential for compressor efficiency improvement and resultant improvement in capacity and system efficiency.
  • Data from the survey tests revealed that based on raw numbers, the Groveport site has the biggest margin between its observed efficiency and the benchmark of 91 to 92%.
  • Operation of the Groveport TCVC10 at a speed of 270 RPM (which is a reduction from the nominal speed for the TCVC of 330 RPM) reduces losses distinctly and increases compressor thermal efficiency.
  • Enthalpy and DIP based efficiencies track quite closely for both sites; enthalpy based efficiency is slightly lower (by 1 to 2 points) than DIP based efficiency.
  • Both candidates have undesirably high pulsations under some operating conditions. The single-acting conditions at Bedford leads to distinctly higher pulsations (a high of over 6% of line pressure) than any other condition tested at either site. For other conditions, the highest pulsations at the two sites are comparable. The control of pulsations, which should accompany any design changes for loss reduction, needs evaluating as part of the planned design studies.
  • Design analysis performed on the HBA-6, using an acoustic model of the suction and discharge piping coupled to the compressor cylinders, predicts significant pulsations at 1X running speed and at the nozzle resonance. The model predictions show that the installation of new bottles would substantially reduce pulsations at running speed and would also reduce nozzle resonance pulsations. The model predicts that installation of a side branch absorber will also reduce pulsations at 1X running speed but will have limited influence on nozzle resonance pulsations. After discussions between SwRI and the host company, the planned modifications for the Bedford site are to install the side branch absorber together with nozzle orifices.

Current Status and Remaining Tasks
Activity remaining under this project includes completion of design analysis for the TCVC10, guided by the test data from the Groveport Station, installation of modifications suggested by design analysis for both Bedford and Groveport stations and follow up testing and analysis at both stations to provide comparative improvement to system operation efficiencies due to the modifications made.

Project Start: September 29, 2002
Project End: February 28, 2006

DOE Contribution: $600,000
Performer Contribution: $442,300

Contact Information:
NETL – Richard Baker (richard.baker@netl.doe.gov or 304-285-4714)
SwRI – Danny Deffenbaugh (ddeffenbaugh@swri.org or 210-522-2384)

Additional Information:
Final Report - May, 2006 Technologies to Enhance Operation of the Existing Natural Gas Compression Infrastructure [PDF-7.25MB]

Project Report December 2005: Manifold Design for Controlling Engine Air Balance [PDF-3827KB] 
 
Project Report May 2005: Compression Infrastructure Project [PDF-1728KB] 
 
Topical Report Oct. 2002 - June 2004: Technologies to Enhance Operation of the Existing Natural Gas Compression Infrastructure  [PDF-5100KB] 
 
Gas Machinery Conference - October, 2004: Enhancing Operation of the Existing Natural Gas Compression Infrastructure - Phase 1 [PDF-1061KB]

StayConnected Facebook Twitter LinkedIn RssFeed YouTube