Brandon Ridens Research Engineer Klaus Brun Program Director Southwest Research Institute. September 9, 2014 Supercritical CO 2 Power Cycle Symposium

Transcription

1 High Pressure Thermophysical Gas Property Testing, Uncertainty Analysis, and Equation of State Comparison for Supercritical CO Compression Applications Brandon Ridens Research Engineer Klaus Brun Program Director Southwest Research Institute September 9, 1 Supercritical CO Power Cycle Symposium

2 Overview Objective Characterization of Representative Gas Mitures Test Description and Methodology Uncertainty Analysis Test Campaign Results and Equation of State Comparisons Conclusions

3 Interest in Supercritical Carbon Dioide Continual increase of government regulations and international standards requiring lower acceptable release limits of greenhouse gases Investigation of sequestration, compression and transportation of CO involving NGCC, IGCC, oy-fuel and pulverized coal power plants Increased interest in compressed supercritical CO Limited data publically available to verify the results from the various equations of state EOS calculations for the range of pressures, temperatures and multi-species gas compositions relevant to compression and pipeline operations

4 Applications to Gas Compression Design and optimization of centrifugal compressor and turbomachinery Density, compressibility, speed of sound, specific heat Mechanical design calculations Relative gas velocity, temperature rise, flow rate for the stage-to-stage impeller geometry, material selection EOS models departure from Ideal Gas Behavior can be profound at etreme pressures and temperatures

5 Representative Gas Mitures and Ranges Miture CO %mol N %mol O %mol Ar %mol CH %mol Description A Membrane Separation B Oy-Comb Lignite-ired C Pipeline CO Mi A Temperature Pressure Temperature Pressure Temperature Pressure Pressure Mi Mi Psia Psia Psia Psia B 15 C *SOS *Density

6 SwRI Gas Property Testing Laboratory: Current Testing Capabilities Density compressibility factor, Speed of Sound and Specific Heat at Constant Volume Cv Stable gas mitures including CO, hydrocarbons, combustibles, non-combustibles and acid gas blends -1, psi pressure range Ambient to 5 temperature range Gas sampling and species determination near critical point gas chromatography Determination of liquid formation Temperature 5 1 Pressure psi 1

7 Density Measurement Based on highly precise mass measurements at constant volume under controlled pressure and temperature conditions High pressure autoclave and precision scale All instruments calibrated by SwRI ISO certified laboratory on site with NIST traceability

8 Speed of Sound Measurement requency measured using an acoustic-to-electric transducer at controlled temperature and pressure conditions Acoustic resonance with a high pressure pipe with two closed ends Eternal speaker used to run a frequency sweep to determine length resonance frequency half-wave response

9 Calculation and Analyses of Uncertainties Primary measurements and reference conditions at each test iteration Direct Measurement Sensor Measurement Uncertainty Scale Precision requency Resolution Geometry Internal Volume Length Reference Condition Pressure Uncertainty Temperature Uncertainty Gas Miture Analytical Uncertainty EOS Model Predictions

10 Application of the Perturbation Method Determination of total uncertainty of dependent measurement systems Does not require linearity assumptions Sequentially perturbing altering input values by their respective uncertainties Use of NIST REPROP program GERG EOS Model = + = = = n i i i i n n n n

11 inal Test Uncertainties Density & SoS Largest component towards total uncertainty gas miture component analytical uncertainty Reference uncertainties that could not be calculated Mi A: 15 psi & 1 Mi B: 18 psi & 1 Mi A Density Measurement Uncertainty Mi C Speed of Sound Measurement Uncertainty Total Uncertainty % Total U Ref state + measured U Measured Density Uncertainty Total Uncertainty % Total U Ref state + measured U Measured Speed of Sound U Measured Density lbm/ft Measured SOS ft/sec

12 Mi A & B Density Results Greater influence of density due to pressure at lower temperatures Larger difference with Mi A Mi A Density Results Mi B Density Results Density Measured kg/m deg deg deg Density Measured kg/m deg deg deg Pressure psi Pressure psi Miture CO %mol O %mol A Miture CO %mol N %mol O %mol B

13 Mi C Density Results Less influence of density due to pressure changes at elevated pressures regardless of temperature. 1 Mi C Density Results 1 Density Measured kg/m deg deg deg Pressure psi Miture CO %mol N %mol Ar %mol CH %mol C

14 Mi A & B Speed of Sound Results Greater influence of speed of sound due to pressure at lower temperatures Impact of the critical point Mi A Speed of Sound Results Mi B Speed of Sound Results Speed of Sound Measured ft/s deg deg deg Speed of Sound Measured ft/s deg deg deg Pressure psi Pressure psi Miture CO %mol O %mol A Miture CO %mol N %mol O %mol B

15 Mi C Speed of Sound Results Similar influence of speed of sound slope due to pressure changes at elevated pressures for all tested temperatures. 1 Mi C Speed of Sound Results Speed of Sound Measured ft/s deg deg deg Pressure psi Miture CO %mol N %mol Ar %mol CH %mol C

16 Mi A Speed of Sound Results Increase of speed of sound when approaching and surpassing the critical point. 16 Mi A Speed of Sound Results Speed of Sound Measured ft/s deg deg deg Pressure psi Miture CO %mol O %mol A

17 Equations of State EOS Comparison GERG 8 Wide-Rage EOS Based on 1 natural gas components Methane, nitrogen, carbon dioide, oygen and argon Default NIST EOS Predict similar density and speed of sound values for all the mitures at most test points Predicted Density Differences A B C.5% kg/m 1.7% kg/m.1% kg/m Predicted Speed of Sound Differences A B C.8% ft/s.% ft/s.19% ft/s

18 Mi A Density EOS Model Comparison Both GERG and Default NIST EOS models match eperimental data within +/- % Lower pressures EOS models predict higher values ma ~6% Higher pressures EOS models predict lower values ma ~6% Similar results for Mi B % Difference Ep - EOS Model Mi A Density Results - EOS Model Comparison Temp = 1 deg selected data 1 Pressure psi % Difference Ep - EOS Model Mi A Density Results - EOS Model Comparison Temp = deg selected data 1 Pressure psia % Difference Ep - EOS Model Mi A Density Results - EOS Model Comparison Temp = deg selected data 1 Pressure psia % Diff GERG % Diff NIST

19 Mi C Density EOS Model Comparison Both GERG and Default NIST EOS models match eperimental data within +/- % for lower temperatures 1 & Larger differences at Higher temperatures EOS models predict lower values ma ~5% Mi C Density Results - EOS Model Comparison Temp = 1 deg selected data Mi C Density Results - EOS Model Comparison Temp = deg selected data Mi C Density Results - EOS Model Comparison Temp = deg selected data % Difference Ep - EOS Model % Difference Ep - EOS Model % Difference Ep - EOS Model % Diff GERG % Diff NIST Pressure psi Pressure psia Pressure psia

20 Mi B SoS EOS Model Comparison Both GERG and Default NIST EOS models match eperimental data within +/- 5% Smaller differences in EOS models and eperimental data at Lower temperatures EOS models predict lower values ma ~6% Similar results for Mi A and Mi C Mi B SOS Results - EOS Model Comparison Temp = 1 deg selected data Mi B SOS Results - EOS Model Comparison Temp = deg selected data Mi B SOS Results - EOS Model Comparison Temp = deg selected data % Difference Ep - EOS Model % Difference Ep - EOS Model % Difference Ep - EOS Model % Diff GERG % Diff NIST Pressure psia Pressure psia Pressure psia

21 EOS Model Conclusions GERG and default NIST EOS models predict similar density and speed of sound values for all the mitures at most test points Ma 1.7% density difference Ma.% speed of sound difference In general, both GERG and default NIST EOS models match eperimental data within +/- 5% for most test points Density pressure influenced Lower pressures EOS models predict higher values Supercritical regime EOS models predict lower values Speed of Sound temperature influenced Smaller differences in EOS models and eperimental data at Lower temperatures EOS models predict lower values

22 Thank You Brandon Ridens Research Engineer Southwest Research Institute