Overview and Interpretation of D7900/D7169 Merge Analysis Crude Oil Quality Association New Orleans, LA March 14, 2019
Value of Merged Simdis Analysis Requires very little sample (10-50 mls) Much faster than Physical Distillation Assay Hours vs Days Quick Feedstock Evaluation (Opportunity crudes, etc.) Ongoing crude quality monitoring
Background D7169 Simulated Distillation Based on technique originally Developed in 1970 s by HP et al Utilizes a boiling point non-polar gas chromatographic column to simulate the mass% boiling point distribution of crude and petroleum products by gas chromatography by correlation of n-paraffin RT and boiling point
Retention Time (RT) vs n-paraffin Boiling Point
RT vs BP(F) Correlation Curve
D7169 Issue 1: LtEnds RT vs BP Correlation Poor C9 Slope ~1 Cannot accurately distinguish BP s from RT s C5
D7169 Issue 2: Poor Resolution of LtEnds Zoomed LtEnds Section of Crude Oil by D7169 HTSD
D7169 Issue 3: Response Quenching by CS2 Solvent CS2 Quenching Effected Region C4 C8 Can Result in Net Lower Than Expected Response up to 2.73 Times Resulting in artificially low recovery!
Issue 3: Response Quenching by CS2 Solvent Excerpt from D7169 Scope
Development of D7900 LtEnds in Crude to C9 Addresses the resolution and BP issues in LtEnds through higher resolution gas chromatographic separation and molecular speciation to improve accuracy
Why Merge D7900 Results To Correct Front End BP Distribution and Recovery of D7169 Assumptions: D7900 LtEnds Mass Recovery and Boiling Points Are More Accurate than D7169 D7900 s use of internal standard for recovery determination more accurate than a generalized external std calibration in D7169 Molecular speciation ensures accurate determination of mass response for C9 and lighter constituents in the GC detector No interference with carbon disulfide causing quench effects Differences in recovery at N-Nonane are absolute, not relative.
Development of D7900 LtEnds in Crude to C9 Accurate LtEnds BP Distribution and Mass Recovery to n-nonane Using Internal Standard Addition for Yield on Whole Crude Better defined Light Ends BP Distribution
Basic Description of Merge Algorithm Collect D7900 Boiling Point mass% data up to the selected cutover point (n-nonane is typical) into result array at a precision of 0.1% or better. Append the D7169 boiling point mass% at the same precision starting at the temperature immediately after the cutover temperature to the result array until all D7169 data is used If mass recovery exceeds 100%, normalize ONLY the D7169 data so that the result is exactly 100% (D7900 deemed accurate as reported) Refer to D7900 Appendix X3 for more details of the algorithm
Graphical Representation of Merging Cutover (n-nonane BP, 303.8 F
Review/Validation of Merge Results Comparing mass recovery at cutover temperature between D7900 and D7169 This difference typically will not exceed 5% in normal stabilized crude oils. If values exceed 5%, individual method results should be reviewed for potential errors: Weighing errors (D7900 and D7169) RT Calibration Old/Inaccurate (D7169) Blank subtraction Using Old or Poor Data (D7169) External standard RF not correct (D7169) Accurate peak identification and calculation of internal standard (D7900,used for yield on whole crude)
Validation of Merge Results Comparing mass recovery at cutover temperature between D7900 and D7169 (cont.) If review indicates no issues with either D7900 or D7169 analyses then: D7900 results should be assumed to be accurate Not affected by solvent quenching or quench factors Significant differences above 5% may occur in crudes containing significant amounts of C6 and lighter. D7169 Quench Factor over/under estimates the amount of light ends Can directly affect the mass recovery at the cutover temperature
D7169 Issue 3: Response Quenching by CS2 Solvent CS2 Quenching Effected Region C4 C8 Can Result in Net Lower Than Expected Response up to 2.73 Times Resulting in artificially low recovery!
Important Concepts When Interpreting Merged BP Results Mass% Differences between D7900 and D7169 at the cutover temperature are absolute, not relative. Differences in recovery shift the distribution parallel to each other
Important Concepts for Interpretation of Merged BP Results There may be results in the merged bp distribution where multiple cumulative percentages occur at the exact same temperature because D7900 is a component-based boiling point distribution as opposed to the slice-based distribution used in Test Method D7169
Other Common Errors in Merging D7900 and D7169 Incorrect Temp Units Merged D7900 In oc, D7169 In of D7169 In oc, D7900 In degf
Other Common Errors in Merging D7900 and D7169 Bi-modality in Heavy Ends D7169 CS2 Wash Runs Should be Performed in Between Crude samples D7169
Comparison of Merged BP Results with TBP Assay Data Merge Crude Oil Simdist Data Is Currently Mass% Data only, Comparisons with TBP data should be on the basis of Mass. New correlations being developed by API may allow for conversion to volume in the near future. It is better practice to compare cutpoints rather than individual mass% recoveries between Merged Simdis and TBP data Compares the average result within a temperature interval, reducing the effect of efficiency differences on results between the two methods. D2892
Comparison of Merged BP Results with TBP Assay Data Vacuum Portion of TBP Distillation D5236 Potstill Change In Efficiency from D2892 D5236
Comparison of Merged BP Results with TBP Cut Assay Data
Conclusions Merge Crude Oil Simulated Distillation is more accurate in terms of mass recovery and Light End BP distribution than D7169 alone The combination of D7900 and D7169 adds valuable information on molecular species not available from HTSD alone Differences in recovery are absolute, not relative; they are not effected by solvent quenching and external standard response variations. When performed correctly, the results are a cost-effective, timeefficient and accurate means of monitoring crude oil quality
Thank You! Question and Answers