Gasoline RFG analysis by a spectrometer

Description

BACKGROUND OF THE INVENTION
I. Field of the Invention
The present invention relates to techniques of analysis, particularly of hydrocarbons and substituted hydrocarbon mixtures generally classified in Class 250.
II. Description of the Prior Art
There are a number of ASTM (American Society for Testing Materials) and other standard methods for the analysis of physical properties of hydrocarbons such as ASTM D-86 distillation temperatures at various percentages distilled; ASTM D-1298 test for API gravity; Gas chromatography --Mass spectrometry (GC Mass Spec) for determination of weight percent aromatics; and ASTM D-2622 for sulfur.
Conventionally, these tests are conducted by enough different test apparatus to fill a small laboratory and are time-consuming and relatively expensive, particularly when applied to the hundreds of samples per day which may be required to be analyzed in a typical refinery. This situation will be made much more acute with the new reformulated fuel requirements for the prediction of total air pollutants (TAPs) exhaust benzene, volatile organic carbon (VOC), nitrogen oxides (NOx), Reid vapor pressure (RVP), and driveability index. Reformulated gasoline is defined in the Federal Register (59CFR32): ". . . any gasoline whose formulation has been certified under .sctn.80.40, which meets each of the standards and requirements prescribed under .sctn.80.41, and which contains less than the maximum concentration of the marker specified in .sctn.80.82 that is allowed for reformulation under .sctn.80.82." Effective Jan. 1, 1995, U.S. goverment regulations require these to be calculated by a "simple" model.
Simple Model
Exhaust Benzene=1.884+0.949(Vol % Benzene)+0.113(Vol % Aromatics-Vol % Benzene)
Total Toxics=Exhaust Benzene+Refueling Benzene+Evaporative Benzene+Running Loss Benzene+Butadiene+Formaldehyde+Acetaldehyde+POM
______________________________________Refueling Benzene = Vol % Benzene (Refueling VOC) (10) [1.3972 - 0.591 (MTBE Wt % Oxygen/2) - 0.081507 (RVP)]Evaporative Benzene = Vol % Benzene (10) (Evaporative VOC) {0.679 [1.4448 - 0.0684 (MTBE Wt % Oxygen/2 - 0.080274 (RVP)]} + 0.0321 [1.3758 - 0.579 (MTBE Wt % Oxygen(2) - 0.080274 (RVP)]Running Loss Benzene = Vol % Benzene (Running loss VOC) (10) [1.4448 - 0.684 (MTBE Wt % Oxygen/2) - 0.080274 (RVP)Butadiene = 0.00556 (Exhaust VOC) 1000Formaldehyde = 0.01256 (Exhaust VOC) 1000 [1 + (0.421/2.7) (MTBE Wt % Oxygen + TAME Wt % Oxygen + (0.358/3.55 (EtOH Wt % Oxygen) + 0.137/2.7) (ETBE Wt % Oxygen + ETAE Wt % Oxygen)]Acetaldehyde = 0.00891 (Exhaust VOC) 1000 [1 + (0.078/2.7) (MTBE Wt % Oxygen + TAME Wt % Oxygen) + (0.865/3.55) (EtOH Wt % Oxygen) + 0.867/2.7) (ETBE Wt % Oxygen + ETAE Wt % Oxygen)]POM = 3.15 (Exhaust VOC)______________________________________
Exhaust VOC=0.444[=0127/2.7)(sum of Wt % Oxygen from MTBE, ETBE, TAME, ETAE)
Refueling VOC=0.04[0.1667(RVP)-0.45]
Evaporative VOC=0.813-0.2393(RVP)+0.21239(RVP)(RVP)
Running Loss VOC=0.2963-0.1306(RVP)+0.016255(RVP)(RVP)
At some time in the future, government regulators may require a more sophisticated "complex" model for prediction of these environmental parameters of fuels. The complex model contains more detailed calculations and two additional parameters: Total NOx (Nitrogen Oxides) and Total VOC (Volatile Organic Carbon). Since individual refineries will have to certify the VOC, NOx, TAP, benzene, and RVP of their fuels on a daily or more frequent basis, the number of tests and their complexity will pose a daunting problem to the refinery industry.
Chemometric models obtained may be used with a laboratory spectrophotometer to determine component concentrations or from physical properties of test samples. Alternatively, an on-line spectrometer installed on or "at" a gasoline stream may be used to predict real time concentrations and physical properties.
A significant amount of work has been done on use of spectroscopy to determine fuel properties.
U.S. Pat. No. 4,963,745 to Maggard teaches octane measured by NIR methyne band, etc.; U.S. Pat. No. 5,223,714 to Maggard teaches a prediction of octane, etc., using linear addition with or without NIR; U.S. Pat. No. 5,349,188 to Maggard teaches the determination of octane by secondary wavelengths in the NIR; U.S. Pat. No. 5,349,189 to Maggard teaches PIANO determination by NIR; U.S. Pat. No. 5,243,546 to Maggard teaches transfer of calibration equations in the NIR; U.S. Pat. No. 5,145,785 to Maggard et al. teaches aromatic content of diesel measured by NIR; U.S. Pat. No. 5,370,790 to Maggard et al. teaches aromatic content of diesel measured by NIR and relates to the apparatus; U.S. Pat. No. 5,348,645 to Maggard et al. teaches the measurement of sulfur content of diesel fuels by NIR; U.S. Pat. No. 5,362,965 to Maggard teaches the direct relationship between the second derivative of the NIR data and concentration of the component.
Experience Leads to Accurate Design of NIR Gasoline Analysis Systems, Welch, Bain, Maggard, and May, Oil & Gas Journal, Jun. 27, 1994, pp 48-56, determines research octane, motor octane, road octane, aromatics, olefins, RVP, benzene, oxygen content, and distillation points during gasoline blending.
U.S. Pat. No. 5,596,196, John B. Cooper et al., teaches Oxygenate Analysis and Control by Raman Spectroscopy.
Reformulated gasoline (RFG) testing thus involves measuring sulfur, olefin, aromatic contents, Reid Vapor Pressure (RVP), and benzene, distillation properties, plus total air pollutants (TAPs), exhaust benzene, volatile organic carbon (VOC), and nitrogen oxides (NOx). Measuring driveability, although not required, is desirable.
All of these tests can be conducted by using a spectrometer, preferably in the IR range, more preferably in the NIR range, and most preferably by a single instrument operating at high-correlation wavelengths. Measured Raman intensities may also be used. Importantly, VOC, TAP, exhaust benzene, NOx, and RVP may be correlated to IR absorbance at certain bands. Statistical methods including partial least squares analysis (PLS), multiple linear regression analysis (MLR), principal component regression analysis (PCR), and neural networks can be used and derivatives of first, particularly second, or other orders can be used. Level 3 or 4 SIMCA may be utilized. As used hereinafter, the term "SIMCA" is employed as commonly understood to refer to soft independent modeling of class analogy. Results can be displayed on a single screen. By predicting simple or complex EPA model values measured by the present invention (NIR, Mid-IR, or Raman embodiments) and comparing the results with results obtained by inserting laboratory values measured by conventional methods into the EPA model, the present invention remains accurate while providing the advantage of labor saving discussed above.
SUMMARY OF THE INVENTION
Accordingly, the present invention provides a process for the determination and/or control of EPA parameters of a gasoline blend (e.g., total toxics, exhaust benzene, VOC's and/or NOx). The present invention comprises in combination, measuring the absorbance or intensity of a liquid hydrocarbon or a component thereof in at least one band in the electromagnetic spectrum, transforming said absorbance by a mathematical transformation comprising multivariate regression analysis, substituting said absorbance or intensity into an equation which predicts one or more EPA parameters or one component thereof of said fuels, and controlling blending of components which affect the TAP of said blend in response to said prediction.
In another embodiment, the present invention provides a method for determining predicted emissions from evaporation and combustion of fuels in an internal combustion engine comprising taking multiple fuel samples and spectrally analyzing each of said samples for at least one of benzene, total aromatics, RVP, wt % oxygen, required distillation points, olefins and sulfur to determine the concentration of said analyzed component using said spectrally determined concentration in a mathematical model for determining total emissions from said determined values and predicting emissions for each fuel sample correlating the spectral data for each fuel sample with said predicted emissions; and predicting emissions from additional fuel samples based on said correlations.

BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a correlation plot of Phase I complex model NOx for winter gasoline, with NIR-predicted NOx plotted against the corresponding Phase I NOx values obtained by inserting the conventional "lab reported" oxygenates, sulfur, RVP, PIANO results for aromatics, olefins and benzene, and distillation by ASTM D-3710 into the EPA complex formula. PIANO is a conventional method which produces results comparable with that of GC-Mass spectrometry for aromatics, fluorescent indicator analysis (FIA) for olefins, and ASDTM D-3606 for benzene by gas chromatography. ASTM D-3710 is a conventional method which produces results comparable with those of ASTM D-86. Oxygenates are determined by ASTM-D5599 and sulfur by ASTM D2622.
FIG. 2 is a similar plot against the same conventional determinations but comparing the Raman embodiment of this invention.
FIG. 3 is a similar plot against conventional determinations but comparing the FT-IR predicted value according to the present invention.
FIGS. 4, 5 and 6 are correlation plots similar to FIGS. 1, 2 and 3, respectively, except that they are for exhaust benzene rather than Phase I NOx
FIGS. 7, 8 and 9 are similar to FIGS. 1, 2 and 3 except that they are for summer gasoline and for Phase II total toxics.
FIGS. 10, 11 and 12 are similar to FIGS. 1, 2 and 3 except that they are correlation plots for summer gasoline and for Phase II total VOC.
FIG. 13 illustrates schematically the fuel blending process described in Example II.
Table I sets forth the methods for boiling point and distillation properties of neat gasolines taken from a set of approximately 110 samples which shows the range (highest and lowest) sample value, the mathematical average, the offset constant, k(0) and constants k(1), k(2) and k(3) which are to be multiplied with the absorbance at wavelengths 1, 2 and 3, respectively. The R-value and R.sup.2 -values are measures of the error and the NIR standard error is actually the standard error of estimates (SE or SEE). The first set of data in Table 1 uses multiple linear regression (MLR) while the second set uses partial least squares (PLS).
Table II is similar to Table I except that it covers MTBE gasolines.
Table III is similar to Table I except that it utilizes Mid-IR.
Table IV is similar to Table I except that it relates to determination of distillation properties of neat gasolines by Raman.
Table V is similar to Table I except that it relates to determination of distillation properties of MTBE gasolines by Raman.
Table VI is similar to Table I except that it relates to determination of EPA parameters for winter gasolines by NIR.
Table VII is similar to Table I except that it relates to determination of EPA parameters of summer gasolines by NIR.
Table VIII is similar to Table I except that it relates to determination of EPA parameters of winter gasolines by Mid-IR.
Table IX is similar to Table I except that it relates to determination of EPA parameters of summer gasolines by Mid-IR.
Table X is similar to Table I except that it relates to determination of EPA parameters for winter gasolines by Raman.
Table XI is similar to Table I except that it relates to determination of EPA parameters for summer gasolines by Raman.
Table XII is similar to Table I except that it covers the driveability index of gasolines of various "gasoline types" (See far left column) by Near-IR, Mid-IR, and Raman spectroscopy (as indicated in the subheadings).
Table XIII is a table relating the error in the primary (conventional) methods for distillation, sulfur, aromatics, olefins, benzene, and oxygenates, to the possible errors for the EPA complex model fuel parameters.
TABLE I__________________________________________________________________________Statistics For Distillation Correlations of Neat Gasolines By__________________________________________________________________________NIRMultiple Linear Regression Equations For Distillation Parameters Primary CONSTANTSProperty Method Range Average K (0) K (1) K (2) K (3)__________________________________________________________________________Initial Boiling Point ASTM D 3710 50.87-130.67 85.662 350.032 -827.077 -491.667 349.31510% ASTM D 3710 66.13-187.24 105.076 756.536 853.319 -886.699 -869.29320% ASTM D 3710 84.66-214.57 130.624 541.348 494.852 -495.969 -946.90430% ASTM D 3710 105.11-238.14 152.861 848.502 877.772 -756.635 710.6550% ASTM D 3710 149.72-278.77 198.504 467.516 -946.696 491.731 524.19770% ASTM D 3710 225.75-314.89 262.184 177.875 -195.956 -132.206 717.69580% ASTM D 3710 259.72-335.49 297.912 246.825 386.051 167.859 261.02290% ASTM D 3710 305.65-377.58 342.029 443.341 -482.256 328.742 294.823End Point ASTM D 3710 371.43-461.74 422.972 345.968 806.877 -660.697 482.001Vol % Dist @ 200 F. ASTM D 3710 14.669-63.226 49.001 -124.272 437.539 185.162 -293.333Vol % Dist @ 300 F. ASTM D 3710 61.755-88.571 79.26 148.001 -259.096 -56.041 -304.617__________________________________________________________________________ Multiple Linear Regression Equations For Distillation Parameters WAVELENGTHS NIR Property 1 2 3 R .sup.. R.sup.2 Std. Error__________________________________________________________________________ Initial Boiling Point 1226 2078 1820 0.8757 0.7669 7.48 10% 1198 1848 2182 0.8555 0.7319 10.6 20% 1828 2080 1228 0.9186 0.8438 9.01 30% 2060 1802 1204 0.9430 0.8892 7.83 50% 1860 1628 1148 0.8964 0.8035 10.8 70% 2154 1642 2060 0.8965 0.8037 10.3 80% 1818 1176 2126 0.8551 0.7312 10.7 90% 1614 1814 1178 0.8026 0.6442 8.17 End Point 1160 1598 1816 0.7057 0.4980 12.2 Vol % Dist @ 200 F. 1238 1804 2066 0.9368 0.8776 3.06 Vol % Dist @ 300 F. 1158 1812 2062 0.9086 0.8256 2.3__________________________________________________________________________PLS Equations For Distillation Parameters PrimaryProperty Method Range Average Wavelength Range (nm)__________________________________________________________________________Initial Boiling Point ASTM D 3710 50.87-130.67 85.415 1214-1264, 2050-2100, 1780-186010% ASTM D 3710 66.13-187.24 105.076 1156-1214, 1780-1860, 2160-220020% ASTM D 3710 84.66-214.57 130.624 1142-1166, 1206-1264, 1660-1670, 1780- 1860, 2050-209430% ASTM D 3710 105.11-238.14 152.861 1144-1242, 1800-1864, 2000-210050% ASTM D 3710 149.72-278.77 198.504 1780-1900, 1600-1670, 1140-121470% ASTM D 3710 225.75-314.89 262.184 1608-1666, 2000-219680% ASTM D 3710 259.72-335.49 297.912 1156-1226, 1780-1856, 2112-216090% ASTM D 3710 305.65-377.58 342.029 1142-1214, 1578-1662, 1790-1860End Point ASTM D 3710 371.43-461.74 422.972 1140-1214, 1560-1630, 1780-1866Vol % Dist @ 200 F. ASTM D 3710 14.669-63.226 49.001 1214-1264, 1780-1860, 2030-2100Vol % Dist @ 300 F. ASTM D 3710 61.755-88.571 79.26 1140-1214, 1780-1860, 2030-2100__________________________________________________________________________ PLS Equations For Distillation Parameters NIR Property Factors R .sup.. R.sup.2 Std. Error__________________________________________________________________________ Initial Boiling Point 3 0.8581 0.7363 7.96 10% 5 0.8764 0.7681 9.91 20% 6 0.9395 0.8827 7.90 30% 7 0.9652 0.9316 6.25 50% 4 0.8840 0.7815 11.49 70% 8 0.9338 0.8720 8.51 80% 11 0.9263 0.8580 8.05 90% 11 0.9043 0.8178 6.04 End Point 10 0.8122 0.6597 10.37 Vol % Dist @ 200 F. 7 0.9484 0.8994 2.82 Vol % Dist @ 300 F. 3 0.9060 0.8208 2.34__________________________________________________________________________ Initial through end points are in degrees Fahrenheit.
TABLE II__________________________________________________________________________Statistics For Distillation Correlations of MTBE Gasolines By__________________________________________________________________________NIRMultiple Linear Regression Equations For Distillation Parameters Primary CONSTANTSProperty Method Range Average K (0) K (1) K (2) K (3)__________________________________________________________________________Initial Boiling Point ASTM D 3710 66.45-105.37 86.246 381.137 -585.282 -805.925 -123.73310% ASTM D 3710 80.24-133.29 105.600 298.181 -1660.32 1279.83 -246.04320% ASTM D 3710 99.45-155.53 127.408 188.89 234.404 794.044 -992.33630% ASTM D 3710 119.67-190.57 146.713 224.173 -336.566 -977.518 248.87550% ASTM D 3710 142.67-277.66 185.149 340.641 674.775 689.043 -600.83570% ASTM D 3710 200.88-310.70 242.015 166.84 -562.526 298.943 270.77380% ASTM D 3710 228.93-337.08 279.224 223.942 -391.06 -569.825 -232.69190% ASTM D 3710 288.61-377.97 334.335 300.678 -122.157 703.648 -150.516End Point ASTM D 3710 384.78-455.14 419.087 613.754 -849.991 -100.541 723.589Vol % Dist @ 200 F. ASTM D 3710 32.166-69.698 54.661 130.875 470.12 776.943 331.167Vol % Dist @ 300 F. ASTM D 3710 63.523-91.000 82.961 5.733 278.857 79.714 149.008__________________________________________________________________________ Multiple Linear Regression Equations For Distillation Parameters WAVELENGTHS NIR Property 1 2 3 R .sup.. R.sup.2 Std. Error__________________________________________________________________________ Initial Boiling Point 1230 1850 2092 0.9251 0.8558 5.66 10% 1228 1940 1390 0.8927 0.7969 8.49 20% 1828 2056 1978 0.8957 0.8023 6.38 30% 1200 1238 1830 0.8986 0.8075 6.82 50% 1178 1214 2040 0.8474 0.7181 12.8 70% 1152 1812 2130 0.9268 0.8590 8.7 80% 1236 2016 2166 0.9084 0.8252 10.6 90% 1794 2052 2164 0.8901 0.7923 7.75 End Point 1228 1642 2046 0.7447 0.5546 12.8 Vol % Dist @ 200 F. 1188 1154 1800 0.9170 0.8409 3.03 Vol % Dist @ 300 F. 1158 1788 2086 0.9419 0.8872 1.63__________________________________________________________________________PLS Equations For Distillation Parameters PrimaryProperty Method Range Average Wavelength Range (nm)__________________________________________________________________________Initial Boiling Point ASTM D 3710 66.45-105.37 86.246 1214-1264, 1780-1860, 2050-216010% ASTM D 3710 80.24-133.29 105.600 1214-1264, 1320-1430, 1900-197020% ASTM D 3710 99.45-155.53 127.408 1780-1860, 1950-210030% ASTM D 3710 119.67-190.57 146.713 1156-1264, 1780-186050% ASTM D 3710 142.67-277.66 185.149 1156-1230, 2000-210070% ASTM D 3710 200.88-310.70 242.015 1140-1214, 1780-1860, 2100-216080% ASTM D 3710 228.93-337.08 279.224 1214-1264, 2000-2040, 2100-220090% ASTM D 3710 288.61-377.97 334.335 1780-1860, 2000-2200End Point ASTM D 3710 384.78-455.14 419.087 1214-1264, 1600-1670, 2000-2100Vol % Dist @ 200 F. ASTM D 3710 32.166-69.698 54.661 1140-1214, 1780-1860Vol % Dist @ 300 F. ASTM D 3710 63.523-91.000 82.961 1140-1214, 1780-1860, 2050-2100__________________________________________________________________________ PLS Equations For Distillation Parameters NIR Property Factors R .sup.. R.sup.2 Std. Error__________________________________________________________________________ Initial Boiling Point 5 0.9291 0.8632 5.59 10% 5 0.9019 0.8134 8.26 20% 6 0.9574 0.9166 4.23 30% 9 0.9602 0.9220 4.54 50% 6 0.9178 0.8424 9.80 70% 4 0.9184 0.8435 9.23 80% 10 0.9625 0.9264 7.29 90% 10 0.9655 0.9322 4.67 End Point 10 0.9136 0.8347 8.20 Vol % Dist @ 200 F. 7 0.9216 0.8493 3.03 Vol % Dist @ 300 F. 6 0.9447 0.8925 1.63__________________________________________________________________________ Initial through end points are in degrees Fahrenheit.
TABLE III__________________________________________________________________________Statistics For Distillation Equations of Neat Gasolines By__________________________________________________________________________Mid-IRMultiple Linear Regression Equations For EPA Parameters Primary CONSTANTSProperty Method Range Average K (0) K (1) K (2) K (3)__________________________________________________________________________Initial Boiling Point ASTM D 3710 50.87-130.67 85.415 115.680 -777.376 1333.380 -2527.99310% ASTM D 3710 66.13-187.24 104.751 88.740 -1784.502 -2286.500 2568.90620% ASTM D 3710 84.66-214.57 130.075 191.305 -1842.247 540.681 1333.96230% ASTM D 3710 105.11-238.14 152.342 164.429 -1398.272 -925.467 1386.21850% ASTM D 3710 149.72-278.77 198.321 144.105 1635.984 -1033.755 1455.08470% ASTM D 3710 225.75-314.89 262.07 274.502 -561.602 1311.308 2027.27080% ASTM D 3710 259.72-335.49 297.588 296.646 619.257 177.189 -1382.25990% ASTM D 3710 305.65-377.58 341.791 331.132 -1526.334 -345.944 -2406.509End Point ASTM D 3710 371.43-461.74 422.382 396.507 1004.513 -3340.712 2824.264Vol % Dist @ 200 F. ASTM D 3710 14.669-63.226 49.162 54.241 438.117 -412.188 336.432Vol % Dist @ 300 F. ASTM D 3710 61.755-88.571 79.28 78.589 147.237 113.544 -728.699__________________________________________________________________________ Multiple Linear Regression Equations For EPA Parameters WAVENUMBERS Mid-IR Property 1 2 3 R .sup.. R.sup.2 Std. Error__________________________________________________________________________ Initial Boiling Point 951.9 2892.4 0.8671 0.7519 7.82 10% 2904 2877 948.1 0.8549 0.7309 10.60 20% 2896.3 1411 959.6 0.8788 0.7723 10.80 30% 2896.3 1488.2 1391.7 0.9012 0.8122 10.10 50% 1565.3 1345.4 867.1 0.8707 0.7581 11.90 70% 2992.7 1399.4 870.9 0.9046 0.8183 9.97 80% 2865.4 1565.3 801.5 0.8594 0.7386 10.70 90% 797.6 747.5 3066 0.8292 0.6876 7.85 End Point 944.2 801.5 805.3 0.7352 0.5405 11.90 Vol % Dist @ 200 F. 1488.2 1395.6 2900.2 0.9054 0.8197 3.73 Vol % Dist @ 300 F. 2900.2 797.6 867 0.9141 0.8356 2.25__________________________________________________________________________PLS Equations For EPA Parameters PrimaryProperty Method Range Average Wavelength Range (cm - 1)__________________________________________________________________________Initial Boiling Point ASTM D 3710 50.87-130.67 85.415 3069.9-2738.110% ASTM D 3710 66.13-187.24 104.751 2931-2842.3, 971.2-743.620% ASTM D 3710 84.66-214.57 130.075 3162.5-2734.3, 2352.3-2240.5, 1654.1-689.630% ASTM D 3710 105.11-238.14 152.342 3147.1-2722.7, 1573-689.650% ASTM D 3710 149.72-278.77 198.321 2904-2900.2, 2348.5-2282.9, 1607.8-1522.9, 1395.6-1303.70% ASTM D 3710 225.75-314.89 262.07 3054.5-2823, 1445.7-1306.6, 936.5-832.380% ASTM D 3710 259.72-335.49 297.588 3093-2726.5, 1781.4-720.490% ASTM D 3710 305.65-377.58 341.791 3093-2726.5, 1781.4-720.4End Point ASTM D 3710 371.43-461.74 422.382 3093-2726.5, 1781.4-720.4Vol % Dist @ 200 F. ASTM D 3710 14.669-63.226 49.162 3093-2726.5, 1781.4-720.4Vol % Dist @ 300 F. ASTM D 3710 61.755-88.571 79.28 3093-2726.5, 1781.4-720.4__________________________________________________________________________ PLS Equations For EPA Parameters Mid-IR Property Factors R .sup.. R.sup.2 Std. Error__________________________________________________________________________ Initial Boiling Point 11 0.887 0.7880 7.488 10% 7 0.9054 0.8197 8.850 20% 9 0.9234 0.8527 8.894 30% 8 0.9384 0.8806 8.218 50% 4 0.9000 0.8100 10.528 70% 10 0.9456 0.8942 7.878 80% 6 0.8633 0.7453 10.705 90% 12 0.9317 0.8681 5.317 End Point 13 0.8986 0.8075 8.052 Vol % Dist @ 200 F. 8 0.9435 0.8902 2.979 Vol % Dist @ 300 F. 7 0.9357 0.8755 1.995__________________________________________________________________________ Initial through end points are in degrees Fahrenheit.
TABLE IV__________________________________________________________________________Statistics For Distillation Equations of NEAT Gasolines By__________________________________________________________________________RAMANMultiple Linear Regression Equations For Distillation Parameters Primary CONSTANTSProperty Method Range Average K (0) K (1) K (2) K (3)__________________________________________________________________________Initial Boiling Point ASTM D 3710 50.87-130.67 85.415 78.651 144.051 205.871 -263.99810% ASTM D 3710 66.13-187.24 104.751 114.843 34.434 324.925 -367.69920% ASTM D 3710 84.66-214.57 130.075 162.660 461.811 -311.060 -180.12130% ASTM D 3710 105.11-238.14 152.342 167.398 463.549 -190.640 -169.18650% ASTM D 3710 149.72-278.77 198.321 206.963 -565.834 -197.504 809.52470% ASTM D 3710 225.75-314.89 262.07 276.240 -13.898 185.143 -209.20380% ASTM D 3710 259.72-335.49 297.588 291.839 -209.878 -89.502 342.89890% ASTM D 3710 305.65-377.58 341.791 326.796 194.670 -182.944 -70.395End Point ASTM D 3710 371.43-461.74 422.382 401.834 -11.951 -193.552 254.718Vol % Dist @ 200 F. ASTM D 3710 14.669-63.226 49.162 56.467 64.089 -138.582 86.031Vol % Dist @ 300 F. ASTM D 3710 61.755-88.571 79.28 85.049 -24.833 84.438 -68.473__________________________________________________________________________ Multiple Linear Regression Equations For Distillation Parameters WAVENUMBERS RAMAN Property 1 2 3 R .sup.. R.sup.2 Std. Error__________________________________________________________________________ Initial Boiling Point 1025.3 898.1 840.2 0.8569 0.7343 8.07 10% 998.3 898.1 840.2 0.8799 0.7742 9.84 20% 1160.3 840.2 797.8 0.9174 0.8416 9.27 30% 1198.9 840.2 797.8 0.9280 0.8612 8.95 50% 385.1 1430.3 254 0.9079 0.8243 10.60 70% 2919.1 1384 793.9 0.9155 0.8381 9.57 80% 813.2 2996.3 1387.9 0.8908 0.7935 9.52 90% 1376.3 813.2 998.3 0.9043 0.8178 5.92 End Point 2919.1 1611.6 1384 0.8381 0.7024 9.50 Vol % Dist @ 200 F. 793.9 516.2 385.1 0.9212 0.8486 3.48 Vol % Dist @ 300 F. 1384 813.2 550.9 0.9291 0.8632 2.10__________________________________________________________________________PLS Equations For Distillation Parameters PrimaryProperty Method Range Average Wavelength Range (cm - 1)__________________________________________________________________________Initial Boiling Point ASTM D 3710 50.87-130.67 85.415 3046.4-2710.8, 2321.9-419.810% ASTM D 3710 66.13-187.24 104.751 3046.4-2710.8, 1681-292.520% ASTM D 3710 84.66-214.57 130.075 1168.1-1137.2, 894.2-751.530% ASTM D 3710 105.11-238.14 152.342 3058-2710.8, 1681-25450% ASTM D 3710 149.72-278.77 198.321 1484.3-223.170% ASTM D 3710 225.75-314.89 262.07 3069.5-2710.8, 1627-230.880% ASTM D 3710 259.72-335.49 297.588 3050.3-2753.3, 1669.5-22790% ASTM D 3710 305.65-377.58 341.791 3096.5-2753.3, 1681-207.7End Point ASTM D 3710 371.43-461.74 422.382 3058-2734, 1681-200Vol % Dist @ 200 F. ASTM D 3710 14.669-63.226 49.162 3096.5-2764.8, 1681-223.1Vol % Dist @ 300 F. ASTM D 3710 61.755-88.571 79.28 3004-2710.8, 1681-200__________________________________________________________________________ PLS Equations For Distillation Parameters RAMAN Property Factors R .sup.. R.sup.2 Std. Error__________________________________________________________________________ Initial Boiling Point 6 0.8961 0.8030 7.037 10% 6 0.9049 0.8188 8.928 20% 4 0.9337 0.8718 8.382 30% 6 0.9495 0.9016 7.633 50% 2 0.8715 0.7595 12.315 70% 6 0.9287 0.8625 8.944 80% 6 0.9050 0.8190 9.030 90% 8 0.9451 0.8932 4.632 End Point 8 0.9094 0.8270 7.407 Vol % Dist @ 200 F. 5 0.9382 0.8802 3.126 Vol % Dist @ 300 F. 6 0.9459 0.8947 1.865__________________________________________________________________________ Initial through end points are in degrees Fahrenheit.
TABLE V__________________________________________________________________________Statistics For Distillation Equations of MTBE Gasolines By__________________________________________________________________________RAMANMultiple Linear Regression Equations For Distillation Parameters Primary CONSTANTSProperty Method Range Average K (0) K (1) K (2) K (3)__________________________________________________________________________Initial Boiling Point ASTM D 3710 66.45-105.37 86.423 117.475 227.230 568.743 -429.49310% ASTM D 3710 80.24-133.29 105.796 123.578 233.993 -583.105 333.39820% ASTM D 3710 99.45-155.53 127.563 181.176 -240.985 -172.335 230.59330% ASTM D 3710 119.67-190.57 146.895 129.841 -202.985 -183.989 431.75850% ASTM D 3710 142.67-277.66 185.528 145.298 -97.209 72.455 419.53970% ASTM D 3710 200.88-310.70 242.031 254.819 499.101 -46.339 352.62180% ASTM D 3710 228.93-337.08 279.094 307.769 -47.342 91.989 589.01790% ASTM D 3710 288.61-377.97 334.185 352.329 -20.561 365.209 -382.598End Point ASTM D 3710 384.78-455.14 419.07 410.292 260.958 -777.788 129.851Vol % Dist @ 200 F. ASTM D 3710 32.166-69.698 54.606 60.008 219.237 -157.971 -91.578Vol % Dist @ 300 F. ASTM D 3710 63.523-91.000 83.011 81.770 4.505 -90.417 27.392__________________________________________________________________________ Multiple Linear Regression Equations For Distillation Parameters WAVENUMBERS RAMAN Property 1 2 3 R .sup.. R.sup.2 Std. Error__________________________________________________________________________ Initial Boiling Point 1021.5 689.8 431.4 0.8959 0.8026 6.61 10% 1021.5 431.4 307.9 0.9098 0.8277 7.96 20% 840.2 797.8 246.2 0.9378 0.8795 5.10 30% 840.2 759.2 311.8 0.9353 0.8748 5.64 50% 2942.3 2911.4 307.9 0.9222 0.8505 9.48 70% 1326.2 2942.3 242.4 0.9392 0.8821 8.20 80% 2942.3 1384 1330 0.9201 0.8466 10.20 90% 2973.1 1372.5 1052.3 0.9213 0.8488 6.73 End Point 1376.3 1052.3 766.9 0.9035 0.8163 8.33 Vol % Dist @ 200 F. 1488.2 311.8 516.2 0.9231 0.8521 2.99 Vol % Dist @ 300 F. 2938.4 1372.5 1009.9 0.9553 0.9126 1.48__________________________________________________________________________PLS Equations For Distillation Parameters PrimaryProperty Method Range Average Wavelength Range (cm - 1)__________________________________________________________________________Initial Boiling Point ASTM D 3710 66.45-105.37 86.423 3038.7-2795.7, 1665.6-20010% ASTM D 3710 80.24-133.29 105.796 2953.8-2734, 1681-20020% ASTM D 3710 99.45-155.53 127.563 3031-2818.8, 1673.3-20030% ASTM D 3710 119.67-190.57 146.895 3096.5-2710.8, 1681-20050% ASTM D 3710 142.67-277.66 185.528 2980.8-2869, 346.5-269.470% ASTM D 3710 200.88-310.70 242.031 2984.7-2926.8, 1403.3-1276.180% ASTM D 3710 228.93-337.08 279.094 3000.1-2903.7, 1422.6-1279.990% ASTM D 3710 288.61-377.97 334.185 3046.4-2710.8, 1681-207.7End Point ASTM D 3710 384.78-455.14 419.07 3096.5-2734, 1673.3-207.7Vol % Dist @ 200 F. ASTM D 3710 32.166-69.698 54.606 3000.1-2710.8, 1657.9-219.2Vol % Dist @ 300 F. ASTM D 3710 63.523-91.000 83.011 3073.4-2726.3, 1673.3-207.7__________________________________________________________________________ PLS Equations For Distillation Parameters RAMAN Property Factors R .sup.. R.sup.2 Std. Error__________________________________________________________________________ Initial Boiling Point 7 0.9115 0.8308 6.3260 10% 7 0.9304 0.8656 7.2641 20% 8 0.9761 0.9528 3.3233 30% 9 0.9846 0.9694 2.9241 50% 5 0.9323 0.8692 9.0030 70% 3 0.9217 0.8495 9.2717 80% 4 0.9083 0.8250 10.9990 90% 9 0.9734 0.9475 4.1645 End Point 9 0.9608 0.9231 5.6621 Vol % Dist @ 200 F. 6 0.9512 0.9048 2.4604 Vol % Dist @ 300 F. 8 0.9754 0.9514 1.1466__________________________________________________________________________ Initial through end points are in degrees Fahrenheit.
TABLE VI__________________________________________________________________________EPA Parameters of Winter Gasolines by NIR__________________________________________________________________________Multiple Linear Regression Equations For EPA Parameters NIR CONSTANTS WAVELENGTHS STDParameter Range Average K (0) K (1) K (2) K (3) 1 2 3 R .sup.. R.sup.2 ERROR__________________________________________________________________________Exhaust 3515-9.403 5.421 1.081 -9.127 -11.517 -19.382 2022 2136 2180 0.9574 0.9166 0.266BenzeneTotal Toxics 18.572-23.981 20.118 21.844 -82.816 -34.182 -8.780 1470 1928 2136 0.9473 0.8974 0.275Exhaust 16.695-57.557 30.740 63.663 264.987 -79.321 -98.160 1430 1800 2132 0.8581 0.7363 3.260BenzeneTotal Toxics 36.694-79.611 52.426 107.888 708.106 -142.487 -110.377 1434 1802 2132 0.8562 0.7331 3.770NO.sub.x 645.836-885.410 749.708 1312.388 -988.749 1289.229 1411.569 1790 1836 2032 0.919 0.8446 22.300Total VOC 578.163-1184.666 680.413 2756.248 -9191.208 -1860.146 1497.272 1462 1654 2062 0.8643 0.7470 46.800Exhaust 34.804-116.108 63.176 220.618 250.674 438.936 -709.153 1178 1834 1908 0.8841 0.7816 5.930BenzeneTotal Toxics 75.756-161.127 107.929 327.050 523.932 -535.690 454.169 1176 1620 1830 0.8824 0.7786 6.880NO.sub.x 1324.24-1843.13 1542.669 1564.034 3408.109 8450.706 -2348.445 1420 1592 1794 0.9188 0.8442 46.600Total VOC 1186.917-2367.174 1377.457 3610.603 93817.390 -3066.116 -45905.830 1574 1652 1462 0.8869 0.7866 82.700__________________________________________________________________________PLS Equations for EPA ParametersParameter Range Average Wavelength Range (nm) Factors R .sup.. R.sup.2__________________________________________________________________________Exhaust Benzene 3.515-9.403 5.421 1156-1214, 1600-1670, 2100-2162 12 0.9840 0.9683 0.172Total Toxics 18.572-23.981 20.118 1156-1214, 1600-1670, 2100-2162 13 0.9825 0.9653 0.168Exhaust Benzene 16.695-57.557 30.740 1132-1214, 1600-1670 12 0.9474 0.8976 2.104Total Toxics 36.694-79.611 52.426 1400-1480, 1600-1670, 1780-1860, 2100-2160 13 0.9425 0.8883 2.534NO.sub.x 645.836-885.410 749.708 1214-1230, 1480-1630, 1780-1860, 2000-2050, 9 0.9568 0.9155 16.849 2100-2200Total VOC 578.163-1184.666 680.413 1142-1148, 1172-1256, 1420-1426, 1594-1670 10 0.9081 0.8246 40.001Exhaust Benzene 34.804-116.108 63.176 1140-1214, 1550-1670, 1780-1860, 1890-1930, 13 0.9410 0.8855 4.466 2100-2200Total Toxics 75.756-161.127 107.929 1156-1214, 1350-1520, 1600-1670, 1950-2020, 11 0.9252 0.8560 5.724 2100-2200NO.sub.x 1324.24-1843.13 1542.669 1156-1214, 1400-1450, 1570-1630, 1780-1860 13 0.9603 0.9222 34.296Total VOC 1186.917-2367.174 1377.457 1138-1252, 1560-1670, 1780-1784 11 0.9270 0.8593 69.202__________________________________________________________________________
TABLE VII - EPA Parameters of Summer Gasolines by NIR Multiple Linear Regression Equations For EPA Parameters CONSTANTS WAVELENGTHS NIR Parameter Range Average K(0) K(1) K(2) K(3) 1 2 3 R R.sup.2 STD ERROR Simple Model Exhaust Benzene 3.928-8.758 6.028 -2.988 -22.506 -46.304 -14.849 2022 2180 2134 0.9345 0.8733 0.396 Total Toxics 20.378-28.947 23.996 11.788 389.099 54.313 -171.058 1138 1634 2134 0.8566 0.7338 1.030 Complex Model Phase 1 Exhaust Benzene 11.621-34.341 22.047 245.214 -3600.977 200.218 -75.836 1584 1834 2134 0.9116 0.8310 2.190 Total Toxics 26.926-55.943 40.762 -72.494 1388.958 973.861 -1154.131 1942 1460 1436 0.8464 0.7164 3.650 NOx 541.621-787.850 667.901 -71.367 3919.725 -531.758 -772.045 1594 1784 2166 0.9413 0.8860 20.600 Total VOC 788.472-1367.134 1017.096 -6476.446 2144.611 6147.0174 18538.080 1652 1812 1854 0.6572 0.4319 90.100 Complex Model Phase 2 Exhaust Benzene 24.038-69.104 45.153 129.761 -898.464 -180.853 500.829 1998 2180 1832 0.8833 0.7802 4.830 Total Toxics 52.163-105.410 77.476 594.559 678.661 -473.665 -905.224 1184 1868 1620 0.8641 0.7467 6.190 NOx 1105.700-1621.780 1363.990 -135.187 -1509.620 -1099.619 8100.105 2166 1784 1594 0.9353 0.8748 45.100 Total VOC 1085.869-1702.454 1294.537 -1757.174 12823.980 -6082.546 -7165.477 1232 1876 2246 0.8467 0.7169 75.800 PLS Equations for EPA Parameters NIR Parameter Range Average Wavelength Range (nm) Factors R R.sup.2 STD ERROR Simple Model Exhaust Benzene 3.928-8.758 6.028 1422-1480 7 0.9494 0.9014 0.371 Total Toxics 20.376-28.947 23.996 1152-1214, 1600-1670, 2100-2160 11 0.9521 0.9065 0.692 Complex Model Phase 1 Exhaust Benzene 11.621-34.341 22.047 1156-1208, 1560-1670 13 0.9756 0.9518 1.523 Total Toxics 26.926-55.943 40.762 1136-1154, 1410-1480, 1600-1670 11 0.9492 0.9010 2.654 NOx 541.621-787.850 667.901 1132-1260, 1320-1430, 1550-1670, 1780-1860 , 2000-220 3 0.9301 0.8651 22.450 Total VOC 788.472-1367.134 1017.096 1132-1202, 1402-1486, 1590-1670, 1780-1860, 2000-223 9 0.9172 0.8413 52.314 Complex Model Phase 2 Exhaust Benzene 24.038-69.104 45.153 1146-1256, 1604-1670 12 0.9813 0.9629 2.301 Total Toxics 52.163-105.410 77.476 1132-1154, 1156-1214, 1502-1626 14 0.9838 0.9679 2.855 NOx 1105.700-1621.780 1363.990 1132-1190, 1592-1670, 1780-1798, 2098-2174 3 0.9339 0.8722 45.611 Total VOC 1085.869-1702.454 1294.537 1132-1152, 1156-1258, 1420-1432 8 0.8715 0.7595 75.2603
TABLE VIII - EPA Parameters of Winter Gasolines by Mid-IR Multiple Linear Regression Equations For EPA Parameters CONSTANTS WAVENUMBERS (cm-1) Mid-IR Parameter Range Average K(0) K(1) K(2) K(3) 1 2 3 R R.sup.2 STD ERROR Simple Model Exhaust Benzene 3.928-8.758 6.028 4.949 89.993 -95.988 13.519 1607.8 1407.2 678 0.9721 0.9450 0.22 Total Toxics 20.378-28.947 23.996 21.273 -70.822 -30.716 -23.077 1005.9 693.5 1503.6 0.9605 0.9226 0.24 Complex Model Phase 1 Exhaust Benzene 11.621-34.341 22.047 31.182 -41.235 -203.784 -276.36 2981.2 1333.9 693.5 0.8781 0.7711 3.51 Total Toxics 26.926-55.943 40.762 41.431 -274.998 -35.574 237.653 693.5 2973.5 1387.9 0.8533 0.7281 4.21 NOx 541.621-787.850 667.901 672.685 -1636.269 5917.339 -3969.414 1580.8 1175.7 894.1 0.8895 0.7912 26.10 Total VOC 788.472-1367.134 1017.096 518.485 -5023.323 5174.315 -3317.536 824.6 940.4 751.3 0.8001 0.6416 52.90 Complex Model Phase 2 Exhaust Benzene 24.038-69.104 45.153 50.899 721.028 -409.136 -627.652 2784.4 1403.3 693.5 0.8839 0.07816 6.86 Total Toxics 52.163-105.410 77.476 102.905 906.374 -430.719 -556.302 940.4 1399.4 693.5 0.8624 0.7437 8.18 NOx 1105.700-1621.780 1363.990 1559.712 -6956.908 16169.63 -3404.095 1569.2 932.6 1399.4 0.8863 0.7855 55.10 Total VOC 1085.869-1702.454 1294.537 1457.897 -4403.561 5490.922 13700.91 793.8 2865.4 1561.5 0.8264 0.6829 95.90 PLS Equations for EPA Parameters NIR Parameter Range Average Wavenumber Range Factors R R.sup.2 STD ERROR Simple Model Exhaust Benzene 3.928-8.758 6.028 3131.6-2803.7, 1584.6-66 6.4 12 0.9861 0.9724 0.16 Total Toxics 20.376-28.947 23.996 3131.6-2769, 1592.3-666.4 10 0.9793 0.9590 0.18 Complex Model Phase 1 Exhaust Benzene 11.621-34.341 22.047 3136-2751, 1657-680 13 0.9571 0.9160 2.22 Total Toxics 26.926-55.943 40.762 1657-680 12 0.9284 0.8619 3.12 NOx 541.621-787.850 667.901 1607.8-666 12 0.946 0.8949 19.25 Total VOC 788.472-1367.134 1017.096 1646.3-666 10 0.8834 0.7804 42.60 Complex Model Phase 2 Exhaust Benzene 24.038-69.104 45.153 1646.3-666 12 0.936 0.8761 5.36 Total Toxics 52.163-105.410 77.476 1646.3-666 12 0.9325 0.8696 6.06 NOx 1105.700-1621.780 1363.990 1646.3-666 12 0.9452 0.8934 40.32 Total VOC 1085.869-1702.454 1294.537 1646.3-666 10 0.8975 0.8055 77.33
TABLE IX - EPA Parameters of Summer Gasolines by Mid-IR Multiple Linear Regression Equations For EPA Parameters CONSTANTS WAVENUMBERS (cm-1) Mid-IR Parameter Range Average K(0) K(1) K(2) K(3) k(4) 1 2 3 4 R R.sup.2 STD ERROR Simple Model Exhaust Benzene 3.928-8.758 6.028 4.008 170.851 -176.626 13.859 1592.3 1291.4 674.2 0.9593 0.9203 0.32 Total Toxics 20.378-28.947 23.996 32.512 -273.19 235.043 50.701 2896.3 2888.6 674.2 0.9276 0.8604 0.98 Complex Model Phase 1 Exhaust Benzene 11.621-34.341 22.047 11.822 -461.329 129.73 306.254 1148.7 836.2 681.9 0.9101 0.8283 2.48 Total Toxics 26.926-55.943 40.762 50.264 -105.264 -283.88 -346.03 1411 824.6 693.5 0.9301 0.8651 2.80 NOx 541.621-787.850 667.901 600.821 -8326.528 6305.637 700.296 1333.9 909.5 670.3 0.9473 0.8974 19.90 Total VOC 788.472-1367.134 1017.096 978.086 0.1635 0.364 -17.8491 -65.958 3104/1138 2992/1200 3108/972 106/850 0.8235 0.6782 79.60 Complex Model Phase 2 Exhaust Benzene 24.038-69.104 45.153 22.023 -796.332 365.58 248.801 1144 817 678 0.922 0.8501 4.50 Total Toxics 52.163-105.410 77.476 53.544 -1004.017 441.743 288.945 1148 817 678 0.9436 0.8904 4.47 NOx 1105.700-1621.780 1363.990 1148.166 -20629.91 14989.53 1772.027 932 910 670 0.9524 0.9071 39.50 Total VOC 1085.869-1702.454 1294.537 892.769 -2.998 -786.624 24.234 0.132 1565/1052 1561/2869 994/1187 824/126 0.9342 0.8727 55.20 PLS Equations for EPA Parameters Mid-IR STD Parameter Range Average Wavenumber Range (cm-1) Factors R R.sup.2 ERROR Simple Model Exhaust Benzene 3.928-8.758 6.028 2776.7-2726.6, 1557.6-14 72.7, 712.7-670.3 2 0.9529 0.9080 0.34 Total Toxics 20.376-28.947 23.996 3127.8-3019.8, 1530.6-1399.4, 840.1-666.4 10 0.9403 0.8842 1.01 Complex Model Phase 1 Exhaust Benzene 11.621-34.341 22.047 2900.2-2857.7, 1210.4-662. 6 4 0.9100 0.8281 2.52 Total Toxics 26.926-55.943 40.762 2826.9-2749.7, 720.5-670.3 8 0.9496 0.9017 2.61 NOx 541.621-787.850 667.901 1430.3-1376.3, 1044.5-874.8 13 0.9879 0.9759 11.72 Total VOC 788.472-1367.134 1017.096 3134-3058, 3008-2962, 1260-1108, 1012-934, 888-812 11 0.9107 0.8294 66.16 Complex Model Phase 2 Exhaust Benzene 24.038-69.104 45.153 3085-2746, 1666-694 8 0.9299 0.8647 7.67 Total Toxics 52.163-105.410 77.476 2900-2865, 820-666 3 0.9253 0.8562 5.12 NOx 1105.700-1621.780 1363.990 2927-2880, 1014-921, 732-666 5 0.9265 0.8584 50.45 Total VOC 1085.869-1702.454 1294.537 3058-2746, 1650-686 11 0.9400 0.8836 60.29
TABLE X - RAMAN EPA Parameters for Winter Gasolines Multiple Linear Regression Equations For EPA Parameters CONSTANTS WAVENUMBERS (cm-1) Mid-IR Parameter Range Average K(0) K(1) K(2) K(3) k(4) 1 2 3 4 R R.sup.2 STD ERROR Simple Model Exhaust Benzene 3.928-8.758 6.028 4.008 170.851 -176.626 13.859 1592.3 1291.4 674.2 0.9593 0.9203 0.32 Total Toxics 20.378-28.947 23.996 32.512 -273.19 235.043 50.701 2896.3 2888.6 674.2 0.9276 0.8604 0.98 Complex Model Phase 1 Exhaust Benzene 11.621-34.341 22.047 11.822 -461.329 129.73 306.254 1148.7 836.2 681.9 0.9101 0.8283 2.48 Total Toxics 26.926-55.943 40.762 50.264 -105.264 -283.88 -346.03 1411 824.6 693.5 0.9301 0.8651 2.80 NOx 541.621-787.850 667.901 600.821 -8326.528 6305.637 700.296 1333.9 909.5 670.3 0.9473 0.8974 19.90 Total VOC 788.472-1367.134 1017.096 978.086 0.1635 0.364 -17.8491 -65.958 3104/1138 2992/1200 3108/972 106/850 0.8235 0.6782 79.60 Complex Model Phase 2 Exhaust Benzene 24.038-69.104 45.153 22.023 -796.332 365.58 248.801 1144 817 678 0.922 0.8501 4.50 Total Toxics 52.163-105.410 77.476 53.544 -1004.017 441.743 288.945 1148 817 678 0.9436 0.8904 4.47 NOx 1105.700-1621.780 1363.990 1148.166 -20629.91 14989.53 1772.027 932 910 670 0.9524 0.9071 39.50 Total VOC 1085.869-1702.454 1294.537 892.769 -2.998 -786.624 24.234 0.132 1565/1052 1561/2869 994/1187 824/126 0.9342 0.8727 55.20 PLS Equations for EPA Parameters Mid-IR STD Parameter Range Average Wavenumber Range (cm-1) Factors R R.sup.2 ERROR Simple Model Exhaust Benzene 3.928-8.758 6.028 2776.7-2726.6, 1557.6-14 72.7, 712.7-670.3 2 0.9529 0.9080 0.34 Total Toxics 20.376-28.947 23.996 3127.8-3019.8, 1530.6-1399.4, 840.1-666.4 10 0.9403 0.8842 1.01 Complex Model Phase 1 Exhaust Benzene 11.621-34.341 22.047 2900.2-2857.7, 1210.4-662. 6 4 0.9100 0.8281 2.52 Total Toxics 26.926-55.943 40.762 2826.9-2749.7, 720.5-670.3 8 0.9496 0.9017 2.61 NOx 541.621-787.850 667.901 1430.3-1376.3, 1044.5-874.8 13 0.9879 0.9759 11.72 Total VOC 788.472-1367.134 1017.096 3134-3058, 3008-2962, 1260-1108, 1012-934, 888-812 11 0.9107 0.8294 66.16 Complex Model Phase 2 Exhaust Benzene 24.038-69.104 45.153 3085-2746, 1666-694 8 0.9299 0.8647 7.67 Total Toxics 52.163-105.410 77.476 2900-2865, 820-666 3 0.9253 0.8562 5.12 NOx 1105.700-1621.780 1363.990 2927-2880, 1014-921, 732-666 5 0.9265 0.8584 50.45 Total VOC 1085.869-1702.454 1294.537 3058-2746, 1650-686 11 0.9400 0.8836 60.29
TABLE XI - RAMAN EPA Parameters for Summer Gasolines Multiple Linear Regression Equations For EPA Parameters CONSTANTS WAVENUMBERS RAMAN Parameter Range Average K(0) K(1) K(2) K(3) 1 2 3 R R.sup.2 STD ERROR Simple Model Exhaust Benzene 3.928-8.758 6.114 6.793 4.766 -3.965 5.582 1615.5 1461.2 994.5 0.9434 0.8900 0.385 Total Toxics 20.376-33.368 24.31 22.562 -24.177 44.348 -21.906 1160.3 990.6 820.9 0.9307 0.8662 0.947 Complex Model Phase 1 Exhaust Benzene 11.621-38.159 22.352 26.763 -198.038 38.479 319.842 412.1 1249.1 608.8 0.9374 0.8787 2.04 Total Toxics 26.926-59.520 41.032 43.044 -196.996 34.126 342.811 412.1 994.5 608.8 0.9525 0.9073 223 NOx 541.621-750.732 661.5 752.811 -93.884 -613.241 809.095 2992.4 1165.6 1349.3 0.9655 0.9322 15.7 Total VOC 788.472-1367.134 1016.812 919.275 12723.91 -1002.92 -4414.41 967.5 763.1 307.9 0.7650 0.5852 87.7 Complex Model Phase 2 Exhaust Benzene 24.038-75.765 45.93 55.976 70.31 624.394 -386.226 1249.1 608.8 412.1 0.9368 0.8776 3.93 Total Toxics 52.163-109.781 77.992 85.748 59.826 605.97 -361.684 994.5 604.9 412.1 0.9407 0.8849 4.37 NOx 1105.700-1539.93 1363.99 1562.041 -202.963 1689.016 -1329.67 2992.4 1665.6 1349.3 0.9663 0.9337 32.3 Total VOC 1085.869-1702.454 1299.398 1401.583 -543.859 4059.328 -2835 1380.2 554.8 331.1 0.8892 0.7907 68.9 PLS Equations for EPA Parameters RAMAN Parameter Range Average Wavenumber Range (cm-1) Factors R R.sup.2 STD ERROR Simple Model Exhaust Benzene 3.928-8.758 6.1114 1623.2-1607.8, 1468.9-1453.5, 1002.2-952.1 3 0.933 0.8705 0 404 Total Toxics 20.376-33.368 24.31 1168.1-1141.1, 1033.1-911.4, 828.6-813.2 7 0.9583 0.9183 0 768 Complex Model Phase 1 Exhaust Benzene 11.621-38.159 22.352 2965.4-2950, 1260.6-1237.5 , 616.5-601.1, 439.1-400.5 7 0.9558 0.9136 1.832 Total Toxics 26.926-59.520 41.032 1002.2-952.1, 635.8-570.2, 450.7-361 .9 7 0.9733 0.9473 1.793 NOx 541.621-750.732 661.5 3023.3-2953.8, 1684.9-1576.9, 1387.9-1310.8 4 0.9386 0.881 21.01 Total VOC 788.472-1367.134 1016.812 1063.9-716.8, 346.5-61.1 12 0.9818 0.9639 30.061 Complex Model Phase 2 Exhaust Benzene 24.038-75.765 45.93 1295.3-1210.5, 647.4-574.1, 450.7-365.8 10 0.9929 0.9859 1.522 Total Toxics 52.163-109.781 77.992 1036.9-955.9, 666.6-547.1, 469.9-358.1 4 0.9178 0.8424 5.279 NOx 1105.700-1539.93 1363.99 3031-2953.8, 1684.9-1607.8, 1387.9-1306.9 7 0.9706 0.9421 32.702 Total VOC 1085.869-1702.454 1299.398 1418.8-1333.9, 593.4-516.2, 369.7-292.5 6 0.921 0.8482 60.8491
TABLE XII__________________________________________________________________________Driveability Index of Gasolines by Near-IR, Mid-IR, and RAMAN Spectroscopies__________________________________________________________________________Multiple Linear Regression Equations For Driveablity Index of Gasolinesby NIR NIR Gasoline CONSTANTS WAVELENGTHS Std.Type Range Average K(0) K(1) K(2) K(3) 1 2 3 R R.sup.2 Error__________________________________________________________________________ Neat 908.180-1472.040 1094.605 2380.663 -3295.76 -1368.48 3917.013 1234 1802 2060 0.9198 0.846 37.6 MTBE 881.665-1408.680 1048.106 1730.997 -1864.35 -7878.87 2750.927 1202 1234 2064 0.9069 0.8225 41.2__________________________________________________________________________PLS Equations For Driveability Index of Gasolines by NIR NIR Gasoline Std. Type Range Average Wavelength Range (nm) Factors R R.sup.2 Error__________________________________________________________________________ Neat 908.180-1472.040 1094.605 1132-1224, 1238-1264, 1594-1670, 11 0.9421 0.8876 33.21 2106-2164 MTBE 881.665-1408.680 1048.106 1170-1228, 1598-1670 10 0.9647 0.9306 27.18__________________________________________________________________________Multiple Linear Regression Equations For Driveability Index of Gasolinesby Mid-IR Mid- IR Gasoline CONSTANTS WAVENUMBERS Std.Type Range Average K(0) K(1) K(2) K(3) 1 2 3 R R.sup.2 Error__________________________________________________________________________ Neat 908.180-1472.040 1094.605 898.48 7570.374 8622.292 -7601.51 952 863.2 1187.3 0.8728 0.7618 47.1__________________________________________________________________________PLS Equations For Driveability Index of Gasolines by Mid-IR NIR Gasoline Std. Type Range Average Wavenumber Range (cm-1) Factors R R.sup.2 Error__________________________________________________________________________ Neat 908.180-1472.040 1094.605 3093-2726.6, 1781-720.5 5 0.8835 0.7806__________________________________________________________________________ 45.62Multiple Linear Regression Equations For Driveability Index of Gasolinesby Raman Ra- man Gasoline CONSTANTS WAVENUMBERS Std.Type Range Average K(0) K(1) K(2) K(3) 1 2 3 R R.sup.2 Error__________________________________________________________________________ Neat 908.180-1472.040 1094.605 954.872 -911.393 -1520.19 3705.512 1476.6 427.5 307.9 0.9274 0.8601 37.3 MTBE 881.665-1408.6 80 1048.106 1109.821 -800.978 -2561.6 2756.446 840.2 388.9 254 0.9053 0.8196__________________________________________________________________________ 42PLS Equations For Driveability Index of Gasolines by Raman Ra- man Gasoline Std. Type Range Average Wavenumber Range (cm-1) Factors R R.sup.2 Error__________________________________________________________________________ Neat 908.180-1472.040 1094.605 3019.4-2710.8, 1681-207.7 8 0.9688 0.9386 25.78 MTBE 881.665-1408.6 80 1048.106 3019.4-2772.6, 1681-207.7 5 0.9179 0.8425__________________________________________________________________________ 39.54
TABLE XIII__________________________________________________________________________Estimated Errors in the Calculations Due to the Repeatabilities, Reproduceabilities, and/or EPA Tolerance Limit for the Primary Methods REPEATABILITY REPRODUCIBILTY EPA BASED POSSIBLE METHOD F(C) F(C) LIMITS UPON EPA TOLERANCE LIMITS ERRORS__________________________________________________________________________DISTILLATIONS IBP ASTM D 3710 2(1) 8(4) WINTER SIMPLE MODEL 10% ASTM D 3710 2(1) 6(3) EXHAUST BENZENE 0.481 TOTAL TOXICS 0.481 20% ASTM D 3710 4(2) 11(6) COMPLEX 1 30% ASTM D 3710 4(2) 13(7) EXHAUST BENZENE 4.875 TOTAL TOXICS 4.171 50% ASTM D 3710 4(2) 13(7) 5 NOx 22.285 70% ASTM D 3710 4(2) 13(7) COMPLEX 2 EXHAUST BENZENE 8.864 80% ASTM D 3710 5(3) 20(11) TOTAL TOXICS 8.248 NOx 49.645 90% ASTM D 3710 7(4) 27(15) 5 SUMMER END PT ASTM D 3710 N/A N/A SIMPLE MODEL EXHAUST BENZENE 0.481 200F ASTM D 3710 2.49 2.5 TOTAL TOXICS 2.051 300F ASTM D 3710 4.2 3.5 COMPLEX 1 EXHAUST BENZENE 3.066 RVP 0.27 0.3 TOTAL TOXICS 4.713 NOx 21.314 COMPOSITION SULFUR ASTM D 2622 0.0021 0.0064 0.0025 COMPLEX 2 EXHAUST BENZENE 6.296 AROMATICS P.I.A.N.O. 2.7 TOTAL TOXICS 7.108 NOx 45.154 OLEFINS P.I.A.N.O. 2.5 BENZENE P.I.A.N.O. 0.21 DRIVEABILITY INDEX 27.5 OXYGENATE ASTM D 5599 mtbe 0.29 0.3 ALL 0.24__________________________________________________________________________

DETAILED DESCRIPTION OF THE INVENTION
Referring to the columns in Table 6, EPA Parameters of Winter Gasolines by NIR, and Table VII, EPA Parameters of Summer Gasolines by NIR; "range" is shows the high and low value measured in a set of roughly 100 samples, "average" s is the mathematical average of the samples for the parameter being measured; k(0) is the offset constant not connected with any particular wavelength; k(1) is a constant which is multiplied by the absorbance at wavelength 1; k(2) is a constant which is multiplied by the absorbance at wavelength 2, etc.; R is the correlation coefficient; and R.sup.2 is the coefficient of determination. Standard error is a measure of the accuracy of calibration (in milligrams per mile) for NIR (Tables VI and VII), for Mid-IR (Tables VIII and IX), and for Raman (Tables X and XI). Similarly, Table XII relates the analogous ranges, averages, constants, and wavelengths for "driveability index" of gasolines by all three methods: NIR, Mid-IR, and Raman.
Tables A-G summarize preferred, more preferred and most preferred parameters of the process, composition and the apparatus of the invention.
TABLE A__________________________________________________________________________High Correlation NIR Spectral Regions for EPA Parameters of WinterGasolines Physical Spectral More Most Property Units Preferred Preferred Preferred__________________________________________________________________________Simple Exhaust nm 800-2500 1132-2300 1132-1156, 1156-1214 Model Benzene 1260-1320, 1600-1670 2000-2100, 2100-2200 Total nm 800-2500 1132-2300 1156-1214, 1430-1510 Toxics 1600-1670, 1900-1970 2100-2162 Complex Exhaust nm 800-2500 1132-2300 1132-1156, 1156-1214 Model Benzene 1320-1430, 1430-1510 Phase I 1600-1670, 1780-1880 1900-1970, 2100-2162 Total nm 800-2500 1132-2300 1132-1156, 1156-1214 Toxics 1400-1480, 1480-1550 1600-1670, 1780-1860 2100-2162 NOx nm 800-2500 1132-2300 1156-1214, 1214-1230 1320-1430, 1480-1600 1600-1670, 1780-1860 2000-2050, 2100-2200 Total VOC nm 800-2500 1132-2300 1132-1156, 1156-1260 1320-1480, 1594-1670 2000-2162 Complex Exhaust nm 800-2500 1132-2300 1132-1156, 1156-1214 Model Benzene 1260-1300, 1350-1520 Phase II 1550-1670, 1890-1930, 1970-2040, 2100-2200 Total nm 800-2500 1132-2300 1156-1214, 1320-1430 Toxics 1430-1520, 1600-1670 1780-1860, 1950-2020 2100-2200 NOx nm 800-2500 1132-2300 1156-1214, 1280-1320 1400-1450, 1570-1600 1600-1670, 1780-1860 2100-2162 Total VOC nm 800-2500 1132-2300 1132-1260, 1320-1480 1550-1670, 1780-1860__________________________________________________________________________
TABLE B__________________________________________________________________________High Correlation NIR Spectral Regions for BPA Parameters of SummerGasolines Physical Spectral More Most Property Units Preferred Preferred Preferred__________________________________________________________________________Simple Exhaust nm 800-2500 1132-230 1132-1156, 1156-1214 Model Benzene 1214-1230, 1230-1264 1264-1320, 1400-1500 1600-1670, 1780-1860 1950-2100, 2100-2200 Total nm 800-2500 1132-230 1132-1156, 1156-1214 Toxics 1400-1480, 1600-1670 2100-2162 Complex Exhaust nm 800-2500 1132-230 1132-1156, 1156-1214 Model Benzene 1430-1480, 1550-1600 Phase I 1600-1670, 1780-1860 2000-2100, 2100-2200 Total nm 800-2500 1132-230 1132-1156, 1156-1214 Toxics 1400-1480, 1600-1670 1900-1970, 2100-2200 NOx nm 800-2500 1132-230 1132-1156, 1156-1214 1214-1230, 1230-1260 1320-1430, 1550-1670 1780-1860, 2000-2200 Total VOC nm 800-2500 1132-230 1132-1214, 1320-1500 1590-1670, 1780-1900 2000-2200 Complex Exhaust nm 800-2500 1132-230 1132-1156, 1156-1214 Model Benzene 1214-1230, 1230-1260 Phase II 1600-1670, 1780-1860 1970-2040, 2100-2200 Total nm 800-2500 1132-230 1132-1154, 1156-1214 Toxics 1450-1600, 1600-1670 1780-1900, 2100-2162 NOx nm 800-2500 1132-230 1132-1156, 1156-1214 1230-1260, 1550-1600 1600-1670, 1780-1860 2070-2200 Total VOC nm 800-2500 1132-2300 1132-1156, 1156-1260 1320-1480, 1780-1900 2160-2300__________________________________________________________________________
TABLE C__________________________________________________________________________High Correlation Mid-IR Spectral Regions for EPA Parameters of WinterGasolines Physical Spectral More Most Property Units Preferred Preferred Preferred__________________________________________________________________________Simple Exhaust cm-1 400-4000 600-1680 3154-2800, 1635-650 Model Benzene 2700-3500 Total cm-1 400-4000 600-1680 3132-2769-1593-650 Toxics 2700-3500 Complex Exhaust cm-1 400-4000 600-1680 3136-2751, 1657-660 Model Benzene 2700-3500 Phase I Total cm-1 400-4000 600-1680 3132-2560, 1657-660 Toxics 2560-3500 NOx cm-1 400-4000 600-1680 3154-2745, 1690-660 2700-3500 Total VOC cm-1 400-4000 600-1680 3132-2734, 1674-666 2700-3500 Complex Exhaust cm-1 400-4000 600-1680 3160-2720, 1647-666 Model Benzene 2700-3500 Phase II Total cm-1 400-4000 600-1680 3140-2730, 1647-650 Toxics 2700-3500 NOx cm-1 400-4000 600-1680 3132-2700, 1684-650 2700-3500 Total VOC cm-1 400-4000 600-1680 3132-2745, 1647-666 2700-3500__________________________________________________________________________
TABLE D__________________________________________________________________________High Correlation Mid-IR Spectral Regions for EPA Parameters of SummerGasolines Physical Spectral More Most Property Units Preferred Preferred Preferred__________________________________________________________________________Simple Exhaust cm-1 400-4000 600-1680 2840-2416, 2416-2366 Model Benzene 2700-3500 1616-1472, 1317-1272 Total cm-1 400-4000 600-1680 3132-2888, 2400-2300 Toxics 2700-3500 1531-1370, 841-650 Complex Exhaust cm-1 400-4000 600-1680 2901-2857, 1650-650 Model Benzene 2700-3500 Phase I Total cm-1 400-4000 600-1680 3140-2749, 1634-608 Toxics 2700-3500 NOx cm-1 400-4000 600-1680 3078-2749, 1696-650 2700-3500 Total VOC cm-1 400-4000 600-1680 3134-2962, 1260-650 2700-3500 Complex Exhaust cm-1 400-4000 600-1680 3166-2746, 1680-650 Model Benzene 2700-3500 Phase II Total cm-1 400-4000 600-1680 3082-2746, 1666-658 Toxics 2700-3500 NOx cm-1 400-4000 600-1680 2932-2746, 1520-1460 2700-3500 1314-1254, 1034-666 Total VOC cm-1 400-4000 600-1680 3093-2746, 1650-608 2700-3500__________________________________________________________________________
TABLE E__________________________________________________________________________High Correlation Raman Spectral Regions for EPA Parameters of WinterGasolines Physical Spectral More Most Property Units Preferred Preferred Preferred__________________________________________________________________________Simple Exhaust cm-1 61-4000 61-1700 3093-2730, 1681-331 Model Benzene 2700-3500 Total cm-1 61-4000 61-1700 3093-2734, 1681-234 Toxics 2700-3500 Complex Exhaust cm-1 61-4000 61-1700 3093-2734, 1681-200 Model Benzene 2700-3500 Phase I Total cm-1 61-4000 61-1700 3093-2726, 1681-200 Toxics 2700-3500 NOx cm-1 61-4000 61-1700 3093-2734, 1666-200 2700-3500 Total VOC cm-1 61-4000 61-1700 3085-2726, 1681-223 2700-3500 Complex Exhaust cm-1 61-4000 61-1700 3093-2726, 1681-200 Model Benzene 2700-3500 Phase II Total cm-1 61-4000 61-1700 3097-2726, 1681-200 Toxics 2700-3500 NOx cm-1 61-4000 61-1700 3097-2718, 1666-200 2700-3500 Total VOC cm-1 61-4000 61-1700 3082-2726, 1674-246 2700-3500__________________________________________________________________________
TABLE F__________________________________________________________________________High Correlation Raman Spectral Regions for EPA Parameters of SummerGasolines Physical Spectral More Most Property Units Preferred Preferred Preferred__________________________________________________________________________Simple Exhaust cm-1 61-4000 61-1700 1624-950 Model Benzene 2700-3500 Total cm-1 61-4000 61-1700 1169-820 Toxics 2700-3500 Complex Exhaust cm-1 61-4000 61-1700 1261-400, 2966-2950 Model Benzene 2700-3500 Phase I Total cm-1 61-4000 61-1700 1003-361 Toxics 2700-3500 NOx cm-1 61-4000 61-1700 3024-2953, 1685-1165 2700-3500 Total VOC cm-1 61-4000 61-1700 1064-61 2700-3500 Complex Exhaust cm-1 61-4000 61-1700 1296-365 Model Benzene 2700-3500 Phase II Total cm-1 61-4000 61-1700 1037-358 Toxics 2700-3500 NOx cm-1 61-4000 61-1700 3031-2953, 1685-1306 2700-3500 Total VOC cm-1 61-4000 61-1700 1419-292 2700-3500__________________________________________________________________________
TABLE G__________________________________________________________________________High Correlation NIR, Mid-IR, and Raman Spectral Regions of Driveability Index of Gasolines Physical Spectral More Most Spectroscopy Property Units Preferred Preferred Preferred__________________________________________________________________________NIR Driveability nm 800-2500 1132-2300 1132-1230, 1230-1264 Index 1594-1670, 1780-1860 1940-2100, 2100-2164 Mid-IR Driveability cm-1 400-4000 600-1781 3093-2726, 1781-720 Index 2700-3500 3500-3300 Raman Driveability cm-1 61-4000 200-1700 3020-2710, 1681-207 Index 2700-3500 3500-3300__________________________________________________________________________
DESCRIPTION OF THE PREFERRED EMBODIMENTS
EXAMPLE I
The Invention Measuring Summer and Winter Gasolines
The purpose of this project is to use near infrared, (NIR, near-IR), mid infrared (mid-IR), and Raman spectroscopies to determine distillation properties, EPA parameters, and the driveability index of finished gasoline by each of near infrared, mid infrared, and Raman spectroscopies. The near-IR spectra are collected using a NIRSystems on-line 5000. The mid-IR spectra are collected using a MIDAC FOx FT-IR. The Raman Spectra are collected using a Nicolet FT-Raman spectrophotometer.
The samples are submitted to gas chromatography for analysis by ASTM D-3710 for distillation parameters. The distillation points of interest are the initial boiling points, 10% recovery, 20% recovery, 30% recovery, 50% recovery, 70% recovery, 80% recovery, 90% recovery, and end point. The volume is distilled at 200.degree. F. and 300.degree. F. and are extrapolated values, using the nearest distillation percentage on each side of the temperature. The samples are divided up into two calibration sets designated neat and MTBE. The neat group includes all non-oxygenated samples, while the MTBE set includes gasolines containing methyl-tert-butyl ether. Multiple linear regression and partial least squares regression are performed on the calibration sets. The results for the neat and MTBE correlations by near-IR are listed in Tables I and II, respectively. The results for the neat correlations by mid-IR are listed in Table III. The results for the neat and MTBE correlations by Raman are listed in Tables IV and V, respectively.
For EPA gasoline exhaust parameters, there are three types of models. There is the "simple model", which came into use effective Jan. 1, 1995, the "complex model phase 1", which goes into effect in 1998, and the "complex model phase 2", which goes into effect in 2000. The samples are divided into two groups because any gasoline with a vapor pressure greater than 8.7 psi is considered winter gasoline, and lower than 8.7 psi is considered summer gasoline. For winter gasolines, the RVP value is set to the fixed value of 8.7 psi in the calculations. However, for summer gasolines, the actual RVP value is used in the calculations. The simple model parameters are calculated using the total aromatics concentration, the benzene concentration, the weight percent oxygen from MTBE, and the Reid vapor pressure. The complex models are calculated using the variables in the simple model, along with the sulfur concentration, the olefins concentration, and the volume percent distilled at 200.degree. F. and 300.degree. F. The results for the winter and summer correlations by near-IR are listed in Tables VI and VII, respectively. The results for the winter and summer correlations by mid-IR are listed in Tables VIII and IX, respectively. The results for the winter and summer correlations by Raman are listed in Tables X and XI, respectively.
With the simple model, the points of interest are the values for exhaust benzene and total toxics (exhaust benzene plus the other volatile organic carbons). With the complex model phase 1 and phase 2, the points of interest are the values for exhaust benzene, total toxics and NOx.
The driveability index is a value for measuring the expected performance during a vehicle cold-start or drive-away. The index is based upon the 10%, 50%, and 90% distillation points of the gasoline. Better performance is expected from those fuels with a lower driveability index value. The gasolines are divided into the same two groups as for the distillation properties. The results for near-IR, mid-IR, and Raman are listed in Table XII.
The repeatability and reproducibility of the primary methods and the EPA tolerance limits are listed in Table XIII. Based upon the EPA tolerance limits, possible errors are calculated by altering the variables used to calculate values of interest in selected samples, and then averaging the differences from the original calculated values. These are also listed in Table XIII.
EXAMPLE II
Invention Controlling a Fuel Blender
FIG. 13 represents a control scheme for an on-line blender in a refinery, with both feed-forward and feedback control loops, utilizing spectral analysis of EPA Fuel Emissions to provide control.
In FIG. 13, the use of multistreaming, whereby the component streams are switched sequentially to a single probe, using valves, is illustrated. However, multiplexing, whereby a probe is located in each component steam and finished gasoline line, or a combination of both, can also be used. In a multistreaming operation such as that illustrated in FIG. 1, component streams 410, 420, 430, 440, 450 and 460 are sequentially routed to the sample cell or sample in line probe of a spectrometer 470 which analyzes each stream for the EPA fuel emissions of interest, e.g., wt % oxygen. An output signal for each stream (proportional to wt % oxygen) is then transmitted to optimizing software such as GINO (Chevron Gasoline Inline Optimization). The GINO software, resident in blending computer 480, then continuously analyzes the signal, optimize and update the blend recipe in response thereto, and downloads the updated recipe to Blend Ratio Control (BRC) software which is resident in Distributed Control System (DCS) 490. The BRC software is capable of controlling DCS 490 which in turn may adjust the position of valves 405, 415, 425, 435, 445, and 455 to change the flow rates of component streams 410, 420, 430, 440, 450 and 460, respectively.
Another spectrometer 500 can also be used in a feedback mode. That is, a slip stream 465 of the finished blend is directed to the sample probe or sample cell of Raman spectrometer 500, which analyzes the finished blend for EPA fuel property compliance and other properties of interest. DCS 490 then receives the feedback signal from spectrometer 500 in the same manner as it receives the feed-forward signals from spectrometer 470. The DCS 490 is configured to allow direct control of valves 405, 415, 425, 435, 445 and 455 by the feedback control loop to override the recipe established by the feed-forward control loop when necessary.
Spectrometer 500 may be the same instrument as spectrometer 470, with feed-forward and feedback functions operating in a multiplexing or multistreaming mode.
Modifications
Specific compositions, methods, or embodiments discussed are intended to be only illustrative of the invention disclosed by this specification. Variation on these compositions, methods, or embodiments are readily apparent to a person of skill in the art based upon the teachings of this specification and are therefore intended to be included as part of the inventions disclosed herein. For example, while the invention has been illustrated with various gasolines, the invention is applicable to diesel fuel. Again, the examples use a simple NIR spectrometer whereas a Fourier Transform near infrared spectrometer, or an FT NIR, or other spectrometer may be substituted.
Reference to documents made in the specification is intended to result in such patents or literature being expressly incorporated herein by reference including any patents or other literature references cited within such documents.

Claims

1. A process for the prediction of an environmental pollution regulatory parameter for a liquid hydrocarbon fuel, due to evaporation and combustion in an internal combustion engine, comprising
a) measuring the absorbance or Raman intensity of the fuel, or of at least one component of the fuel, with a spectrometer in at least one band of the electromagnetic spectrum;
b) transforming the absorbance or Raman intensity measured in step a) by mathematical transformation comprising multivariant regression analysis to obtain a mathematically transformed absorbance or Raman intensity value;
c) substituting said transformed absorbance or Raman intensity value into an equation or correlation which predicts said environmental pollution regulatory parameter, or one or more input values to a model for obtaining said environmental pollution regulatory parameter, of a liquid hydrocarbon fuel; and
d) obtaining a prediction of said environmental pollution regulatory parameter or one or more of said input values of said liquid hydrocarbon fuel.
2. The process of claim 1 in which the environmental pollution regulatory parameter is selected from total toxics, exhaust benzene, volatile organic carbon, nitrogen oxides, and Reid vapor pressure.
3. The process of claim 1 in which the mathematical transformation includes use of a derivative.
4. The process of claim 1 in which the mathematical transformation comprises multiple linear regression, partial least squares, principal component regression, or level 3 or level 4 SIMCA and/or neural network.
5. The process of claim 4 in which the absorbance is measured and is accomplished in the near-infrared range or the mid-infrared range, and the fuel is gasoline.
6. The process of claim 5 in which the environmental pollution regulatory parameter is selected from total toxics, exhaust benzene, volatile organic carbon, nitrogen oxides, and Reid vapor pressure.
7. The process of claim 1 in which the absorbance is measured and is accomplished in the near-infrared range or the mid-infrared range, and the fuel is gasoline.
8. The process of claim 1 in which the Raman intensity is measured and the measuring of the Raman intensity is accomplished by a Raman Spectrometer, and the fuel is gasoline.
9. A process for controlling a fuel blending process to provide a liquid hydrocarbon fuel for an internal combustion engine in conformance with environmental pollution regulatory parameters comprising
a) measuring the absorbance or Raman intensity of a fuel, or at least one component thereof, with a spectrometer in at least one band of the electromagnetic spectrum;
b) transforming the absorbance or Raman intensity measured in step a) by mathematical transformation comprising multivariant regression analysis to obtain a mathematically transformed absorbance or Raman intensity value;
c) substituting said transformed absorbance or Raman intensity value into an equation or correlation which predicts said environmental pollution regulatory parameters, or one or more input values to a model for obtaining said environmental pollution regulatory parameters, of a fuel, and obtaining a predicted environmental pollution regulatory parameter or one or more of said input values of the liquid hydrocarbon fuel; and
d) utilizing the predicted environmental pollution regulatory parameter or one or more of said input values from step c) in controlling a fuel blending process.
10. The process of claim 9 in which the parameter is selected from total toxics, exhaust benzene; volatile organic carbon; nitrogen oxides, and Reid vapor pressure.
11. The process of claim 9 in which the mathematical transformation comprises use of a derivative.
12. The process of claim 9 in which the mathematical transformation comprises multiple linear regression, partial least squares, principal component regression, or level 3 or level 4 SIMCA and/or neural network.
13. The process of claim 12 in which the absorbance is measured and is accomplished in the near-infrared range or the mid-infrared range, and the fuel is gasoline.
14. The process of claim 13 in which the environmental pollution regulatory parameter is selected from total toxics, exhaust benzene, volatile organic carbon, nitrogen oxides, and Reid vapor pressure.
15. The process of claim 12 in which the absorbance is measured and is accomplished in the near-infrared range or the mid-infrared range, and the fuel is a diesel fuel.
16. The process of claim 9 in which the absorbance is measured and is accomplished in the near-infrared range or the mid-infrared range, and the fuel is gasoline.
17. The process of claim 9 in which the Raman intensity is measured and the measuring of the Raman intensity is accomplished by a Raman Spectrometer, and the fuel is gasoline.
18. A method for predicting emissions from evaporation and combustion of fuel in an internal combustion engine comprising
a) taking multiple fuel samples and spectrally analyzing each of said samples to determine the concentration for at least one of benzene, total aromatics, weight percent oxygen, olefins, and sulfur;
b) utilizing at least one concentration determined in step a) in a mathematical model or correlation for predicting total emissions from a concentration or concentrations determined in step a), and obtaining predicted emissions for each fuel sample;
c) correlating spectral data obtained for each fuel sample with predicted emissions for each sample to obtain correlations between the spectral data and the predicted emissions;
d) obtaining spectral data of additional fuel samples and predicting emissions for each of the additional fuel samples based on the correlations obtained in step c).
19. The method of claim 18 in which the fuel is gasoline, the fuel samples are spectrally analyzed using absorbance measurement, and absorbance measurement is accomplished in the near infrared range or the mid-infrared range.
20. A method for predicting emissions from evaporation and combustion of fuel in an internal combustion engine comprising
a) taking multiple fuel samples and spectrally analyzing each of said samples to determine the Reid vapor pressure and/or distillation points thereof;
b) utilizing at least one of the Reid vapor pressure and/or distillation points determined in step a) in a mathematical model or correlation for predicting total emissions from Reid vapor pressure and/or distillation points determined in step a), and obtaining predicted emissions for each fuel sample;
c) correlating spectral data obtained for each fuel sample with predicted emissions for each sample to obtain correlations between the spectral data and the predicted emissions;
d) obtaining spectral data of additional fuel samples and predicting emissions for each of the additional fuel samples based on the correlations obtained in step c).
21. The method of claim 20 in which the fuel is gasoline, the fuel samples are spectrally analyzed using absorbance measurement, and absorbance measurement is accomplished in the near infrared range or the mid-infrared range.
22. A method for controlling fuel blending for a fuel for an internal combustion engine comprising
a) taking multiple fuel samples and spectrally analyzing each of said samples to determine the concentration for at least one of benzene, total aromatics, weight percent oxygen, olefins, and sulfur;
b) utilizing at least one concentration determined in step a) in a mathematical model or correlation for predicting total emissions from a concentration or concentrations determined in step a), and obtaining predicted emissions for each fuel sample;
c) correlating spectral data obtained for each fuel sample with predicted emissions for each sample to obtain correlations between the spectral data and the predicted emissions;
d) obtaining spectral data for an additional fuel sample and predicting emissions for the additional fuel sample based on the correlations obtained in step c); and
e) controlling a fuel blending process utilizing the correlations and predicting of emissions obtained in steps c) and d).
23. The method of claim 22 in which the fuel is gasoline, the fuel samples are spectrally analyzed using absorbance measurement, and absorbance measurement is accomplished in the near infrared range or the mid-infrared range.
24. The method of claim 22 in which the fuel is a diesel fuel, the fuel samples are spectrally analyzed using absorbance measurement, and absorbance measurement is accomplished in the near infrared range or the mid-infrared range.
25. A method for controlling fuel blending for a fuel for an internal combustion engine comprising
a) taking multiple fuel samples and spectrally analyzing each of said samples to determine the Reid vapor pressure and/or distillation points thereof;
b) utilizing at least one of the Reid vapor pressure and/or distillation points determined in step a) in a mathematical model or correlation for predicting total emissions from Reid vapor pressure and/or distillation points determined in step a), and obtaining predicted emissions for each fuel sample;
c) correlating spectral data obtained for each fuel sample with predicted emissions for each sample to obtain correlations between the spectral data and the predicted emissions;
d) obtaining spectral data for an additional fuel sample and predicting emissions for the additional fuel sample based on the correlations obtained in step c); and
e) controlling a fuel blending process utilizing the correlations and predicting of emissions obtained in steps c) and d).
26. The method of claim 25 in which the fuel is gasoline, the fuel samples are spectrally analyzed using absorbance measurement, and absorbance measurement is accomplished in the near infrared range or the mid-infrared range.
27. A process for controlling a fuel blending process to provide a liquid hydrocarbon fuel for an internal combustion engine in conformance with environmental pollution regulatory parameters comprising
a) measuring the absorbance or Raman intensity of a fuel, or at least one component thereof, with a spectrometer in at least one band of the electromagnetic spectrum;
b) substituting said absorbance or Raman intensity into a correlation which predicts said environmental pollution regulatory parameter, or one or more input values to a correlation for obtaining said environmental pollution regulatory parameters, of a fuel, and obtaining a predicted environmental pollution regulatory parameter or one or more of said input values of the liquid hydrocarbon fuel; and
c) utilizing the predicted environmental pollution regulatory parameter or one or more of said input values from step b) in controlling a fuel blending process.
28. The process of claim 27 in which the fuel is gasoline and the environmental pollution regulatory parameter is selected from total toxics, exhaust benzene, volatile organic carbon, nitrogen oxides, and Reid vapor pressure.
29. A process comprising
a) measuring the absorbance or Raman intensity of a fuel, or at least one component of the fuel, with a spectrometer in at least one band of the electromagnetic spectrum;
b) transforming the absorbance or Raman intensity measured in step a) by mathematical transformation comprising multivariant regression analysis to obtain a mathematically transformed absorbance or Raman intensity value;
c) substituting said transformed absorbance or Raman intensity value into a correlation which predicts said environmental pollution regulatory parameter, or one or more input values to a model for obtaining said environmental pollution regulatory parameter, of a liquid hydrocarbon fuel; and
d) predicting emissions and/or one or more input values of the fuel upon substituting the transformed absorbance or Raman intensity value into said correlation, the correlation being the simple Environmental Protection Agency Model or a complex Environmental Protection Agency Model.
30. A process for determining environmental pollution regulatory parameters due to evaporation and combustion of a fuel for an internal combustion engine comprising
a) measuring, with a spectrometer, at at least one band of the electromagnetic spectrum, the absorbance or Raman intensity of a calibration sample of a liquid hydrocarbon fuel, or component of the fuel;
b) performing mathematical transformation on the absorbance or Raman intensity measured, or a function thereof, of said calibration sample as individual independent variables in a model;
c) assigning and applying weighting constants, or their equivalents, to said independent variables;
d) calibrating a spectrometer utilizing known environmental pollution regulatory parameters and weighted individual independent variables produced from steps b) and c);
e) measuring, with the spectrometer calibrated, at at least one band of the electromagnetic spectrum, the absorbance or Raman intensity of a different liquid hydrocarbon fuel, or component thereof, of undetermined environmental pollution regulatory parameters;
f) performing mathematical transformation on the absorbance or Raman intensity measured, or a function thereof, as individual independent variables in a model, of said different liquid hydrocarbon fuel, or component thereof, of undetermined environmental pollution regulatory parameters;
g) applying the weighting constants, or equivalents thereof, to the individual independent variables obtained in step f) to determine one or more of the environmental pollution regulatory parameters of the different liquid hydrocarbon fuel.
31. The process of claim 30 in which the mathematical transformation includes use of a derivative and comprises multiple linear regression, partial least squares, principal component regression, or level 3 or level 4 SIMCA and/or neural network.
32. The process of claim 31 wherein the fuel is gasoline.
33. A process for control of environmental pollution regulatory parameters due to evaporation and combustion of a fuel for an internal combustion engine comprising
a) measuring, with a spectrometer, at at least one band of the electromagnetic spectrum, the absorbance or Raman intensity of a calibration sample of a liquid hydrocarbon fuel, or component of the fuel;
b) performing mathematical transformation on the absorbance or Raman intensity measured, or a function thereof, of said calibration sample as individual independent variables in a model;
c) assigning and applying weighting constants, or their equivalents, to said independent variables;
d) calibrating a spectrometer utilizing known environmental pollution regulatory parameters and weighted individual independent variables produced from steps b) and c);
e) measuring, with the spectrometer calibrated, at at least one band of the electromagnetic spectrum, the absorbance or Raman intensity of a different liquid hydrocarbon fuel, or component thereof, of undetermined environmental pollution regulatory parameters;
f) performing mathematical transformation on the absorbance or Raman intensity measured, or a function thereof, as individual independent variables in a model, of said different liquid hydrocarbon fuel, or component thereof, of undetermined environmental pollution regulatory parameters;
g) applying the weighting constants, or equivalents thereof, to the individual independent variables obtained in step f) to determine one or more of the environmental pollution regulatory parameters of the different liquid hydrocarbon fuel; and
h) controlling environmental pollution regulatory parameters, related to evaporation and combustion, of a fuel in response to values of the parameters determined in step g).
34. The process of claim 33 in which the mathematical transformation comprises multiple linear regression, partial least squares, principal component regression, or level 3 or level 4 SIMCA and/or neural network, the absorbance is measured, and the measuring is accomplished in the near-infrared range or the mid-infrared range.
35. The process of claim 34 in which the absorbance is measured and is accomplished in the near-infrared range or the mid-infrared range, and the fuel is gasoline.
36. The process of clam 34 in which the fuel is a diesel fuel.
37. The process of claim 33 in which the absorbance is measured and is accomplished in the near-infrared range or the mid-infrared range, and the fuel is gasoline.
38. A process for determining environmental pollution regulatory parameters due to evaporation and combustion of a fuel for an internal combustion engine comprising
a) measuring, with a spectrometer, at at least one band of the electromagnetic spectrum, the absorbance or Raman intensity of a calibration sample of a liquid hydrocarbon fuel, or component of the fuel;
b) transforming the absorbance or Raman intensity measured in step a) by mathematical transformation comprising multivariant regression analysis to obtain a mathematically transformed absorbance or Raman intensity value;
c) substituting said transformed absorbance or Raman intensity value into an equation or correlation which predicts said environmental pollution regulatory parameter, or one or more input values to a model for obtaining said environmental pollution regulatory parameter, of a liquid hydrocarbon fuel, and obtaining a predicted environmental pollution regulatory parameter of the liquid hydrocarbon fuel;
d) calibrating a spectrometer utilizing known environmental pollution regulatory parameters and the environmental pollution regulatory parameter predicted in step c);
e) measuring, with the spectrometer calibrated, at at least one band of the electromagnetic spectrum, the absorbance or Raman intensity of a different liquid hydrocarbon fuel, or component thereof, of undetermined environmental pollution regulatory parameters;
f) transforming the absorbance or Raman intensity measured of said undetermined liquid hydrocarbon fuel by mathematical transformation comprising multivariant regression analysis to obtain a mathematically transformed absorbance or Raman intensity value;
g) substituting said transformed absorbance or Raman intensity value into an equation or correlation which predicts an environmental pollution regulatory parameter, or one or more input values to a model for obtaining an environmental pollution regulatory parameter, of a fuel, and obtaining a predicted environmental pollution regulatory parameter of said different liquid hydrocarbon fuel.
39. The process of claim 38 in which the mathematical transformation includes use of a derivative and comprises multiple linear regression, partial least squares, principal component regression, or level 3 or level 4 SIMCA and/or neural network.
40. The process of claim 39 in which the absorbance is measured and is accomplished in the near-infrared range or the mid-infrared range, and the fuel is gasoline.
41. The process of claim 38 in which the absorbance is measured and is accomplished in the near-infrared range or the mid-infrared range, and the fuel is gasoline.
42. The process of claim 38 in which the fuel is a diesel fuel.
43. A process for control of environmental pollution regulatory parameters due to evaporation and combustion of a fuel for an internal combustion engine comprising
a) measuring, with a spectrometer, at at least one band of the electromagnetic spectrum, the absorbance or Raman intensity of a calibration sample of a liquid hydrocarbon fuel, or component of the fuel;
b) transforming the absorbance or Raman intensity measured in step a) by mathematical transformation comprising multivariant regression analysis to obtain a mathematically transformed absorbance or Raman intensity value;
c) substituting said transformed absorbance or Raman intensity value into an equation or correlation which predicts said environmental pollution regulatory parameter, or one or more input values to a model for obtaining said environmental pollution regulatory parameter, of a liquid hydrocarbon fuel, and obtaining a predicted environmental pollution regulatory parameter or one or more of said input values of the liquid hydrocarbon fuel;
d) calibrating a spectrometer utilizing known environmental pollution regulatory parameters and the environmental pollution regulatory parameter or one or more of said input values predicted in step c);
e) measuring, with the spectrometer calibrated, at at least one band of the electromagnetic spectrum, the absorbance or Raman intensity of a different liquid hydrocarbon fuel, or component thereof, of undetermined environmental pollution regulatory parameters;
f) transforming the absorbance or Raman intensity measured of said undetermined liquid hydrocarbon fuel by mathematical transformation comprising multivariant regression analysis to obtain a mathematically transformed absorbance or Raman intensity value;
g) substituting said transformed absorbance or Raman intensity value into an equation or correlation which predicts an environmental pollution regulatory parameter, or one or more input values to a model for obtaining an environmental pollution regulatory parameter, of a fuel, and obtaining a predicted environmental pollution regulatory parameter or one or more of said input values of said different liquid hydrocarbon fuel; and
h) controlling environmental pollution regulatory parameters due to evaporation and combustion of a fuel in response to the environmental pollution regulatory parameter or one or more of said input values predicted in step g).
44. The process of claim 43 in which the mathematical transformation comprises multiple linear regression, partial least squares, principal component regression, or level 3 or level 4 SIMCA and/or neural network, the absorbance is measured, and the measuring is accomplished in the near-infrared range or the mid-infrared range.
45. The process of claim 44 in which the absorbance is measured and is accomplished in the near-infrared range or the mid-infrared range, and the fuel is gasoline.
46. The process of claim 43 in which the absorbance is measured and is accomplished in the near-infrared range or the mid-infrared range, and the fuel is gasoline.

US Referenced Citations (8)

Number	Name	Date
5223714	Maggard	Jun 1993
5349188	Maggard	Sep 1994
5349189	Maggard	Sep 1994
5360972	DiFoggio et al.	Nov 1994
5412581	Tackett	May 1995
5475612	Espinosa et al.	Dec 1995
5596196	Cooper et al.	Jan 1997
5712481	Welch et al.	Jan 1998

Gasoline RFG analysis by a spectrometer

Information

Patent Number

Date Filed

Date Issued

Inventors

Original Assignees

Examiners

Agents

CPC

US Classifications

Field of Search

US

International Classifications

Abstract

Description

Claims

US Referenced Citations (8)