US3121677A

US3121677A - Process for controlling carbon residue content of oil

Info

Publication number: US3121677A
Application number: US65778A
Authority: US
Inventors: Norman D Coggeshall; Matthew S Norris
Original assignee: Gulf Research and Development Co
Current assignee: Gulf Research and Development Co
Priority date: 1960-10-28
Filing date: 1960-10-28
Publication date: 1964-02-18
Anticipated expiration: 1981-02-18

Description

18, 1954 N. D. COGGESHALL ETAL 3,121,677

PROCESS FOR CONTROLLING CARBON RESIDUE CONTENT OF OIL Filed Oct. 28, 1960 V 3 Sheets-Sheet 1 SCALE UNITS O I l I I I O 0.2 0.4 0.6 0.8 L0 L2 L4 L6 CARBON RESIDUE (RAMSBOTTOM) INVENTORS F NORMAN D. COGGESHALL MATTHEW S. NORRIS BY ATTORNEY 13, 1964 N. D. COGGESHALL ETAL 3,121,677

PROCESS FOR CONTROLLING CARBON RESIDUE CONTENT OF OIL Filed Oct. 28, 1960 s Sheets-Sheet 2 O O O O O O O G) (0 1 (\1 w: 1x19 asm 0017-0229 vaav NORMAN D. COGGESHALL B[{jATTHEW S. NORRIS ATTORNEY INVENTORS United States Patent aware Filed Get. 28, 1969, Ser. No. 65,778 12 Ciainrs. (Cl. 268-173) This invention relates to a rapid method for quantitatively determining the presence in hydrocarbon oils of materials that are indicative of the carbon residue content of such oils, and particularly to the use of such method as a basis for the control of petroleum refining operations.

The carbon residue content of hydrocarbon oils is used frequently in petroleum refining or process operations as an index of the quality of such oils. The term carbon residue is defined as the carbonaceous residue formed after evaporation and pyrolysis of a petroleum product under controlled conditions. The residue is not entirely composed of carbon but is a coke that can be further changed by pyrolysis. Carbon residue values arenormally reported in terms of percent by Weight on the oil sample subjected to the test.

As an example of the use of the carbon residue content of an oil as an index of the quality of an oil, the carbon residue of the side stream of a vacuum distillation unit in which is prepared the heavy gas oil portion of the feed to a catalytic cracking unit can be used as an index of the amount of coke that will be formed by the feed on the cracking catalyst. The carbon residue of the combined feed to a catalytic cracking unit also can be sed in the same way. Similarly, the carbon residue of various other hydrocarbon oil processing streams, for example, the products of primary or finishing distillation operations carried out on lubricating oil stocks, or the product streams obtained from the solvent treating of lubricating oil stocks can be used as an index of the efficiency of the particular processing operation in question. In each case, in the event of deviation from a preselected optimum carbon residue value, one or more appropriate processing variables can be modified in such a Way as to change the carbon residue content of the product stream to the desired level, or alternatively, to compensate for the particular carbon residue content found.

Heretofore it has been necessary from a practical standpoint to determine carbon residues by one or both of two laboratory methods, the so-called Conradson method (ASTM D489) and the Ramsbottom method (ASTM D-524). Each of these methods involves essentially evaporating and partially coking under control conditions the oil to be analyzed, Weighing the residue, and determining the proportion of the original sample represented by the residue. Each method normally requires at least about tWo hours to carry out. As such, neither method is adapted for close monitoring and process control. By the time oil samples are taken, analyses run, and results returned to the operator, many barrels of product having too high or too low a carbon residue may have been produced by the unit. As a consequence, it frequently has been the practice to carry ice out petroleum refining operations that are dependent on carbon residue determinations, not at the optimum level, but rather at carbon residue levels that are well. within the maximum permissible or optimum limits. This practice is not entirely satisfactory for the reason that the carbon residue content is maintained at a safe level only with a proportionate sacrifice of a portion of the potential yield of upgraded products that would be obtainable if the particular process in question were operated at or above the maximum permissible carbon residue level.

The present invention relates to controlling a hydrocarbon oil refining process in response to changes in the carbon residue content of an oil stream in such process. It has now been found that certain components of a hydrocarbon oil that are indicative of the carbon residue content of the oil can be rapidly determined quantitatively by spectrometric methods, whereby oil streams in refining operations can be closely monitored for changes in carbon residue and whereby improved yields of upgraded products can be obtained from such operations without undue sacrifice in the quality of such product. In accordance With the present invention a plurality of oil samples taken at different times from the hydrocarbon oil stream to be analyzed are subjected to spectrometric analysis by passing light therethrough of a wavelength that is selectively absorbable by the components of the oil that are indicative of carbon residue content, in a proportion related to the quantity of such components in the oil samples. Excellent results have been obtained by the use of light radiation having a plurality of Wavelengths that are at least representative of a wavelength band of about 350 to 400 millimicrons, but the invention is not limited to the use of light of this broad wavelength band, as good results have also been obtained with light radiation having a wavelength of only about 400 millimicrons. The radiation that is not absorbed by the oil samples, that is, the transmitted light, is electrically detected and converted to output signals Whose intensity is related to the amount of radiation absorbed by the oil samples, and consequently, under Beers law, to the concentration in the oil samples of the components that are indicative of carbon residue content. In connection with the use of light of wavelengths representative of the wavelength band of about 350 to 400 millimicrons, the integrated absorptivity in the indicated Wavelength band, that is, the area under the absorptivity curve: over the Wavelength band indicated, has been found to coorelate with carbon residue in excellent fashion. However, the present invention is not limited to the use of the integrated absorptivity as an index of carbon residue content, as other functions of the absorption signal intensity that are related to the concentration in the oil of components that are indicative of carbon residue can be used. For example, good correlation With carbon residue can also be obtained from the integrated absorbance in the 350 to 400 millimicron Wavelength band, or directly from the absorptivity, the absorbance, or indeed, the absorption signal intensity itself, with light of a single Wavelength of about 400 millicrons. For purposes of hydrocarbon oil refining process control, one or more variable conditions of a selected operation of the process, for example, the conditions under which the oil stream is prepared, e.g., distillation conditions, or alternatively, the conditions of a process operation to which the oil stream undergoing analysis is to be conducted, e.g., a catalytic cracking operation, are adjusted in response to changes in the intensity of the absorption signal or a function thereof that is also related to the concentration in the oil samples of the components that are indicative of carbon residue, so as to correct the carbon residue content of the stream to a preselected level, or alternatively, to compensate for the change in carbon residue. Although the invention is especially adapted for controlling hydrocarbon oil processing operations, it involves not only the method for controlling such processes, but also various subcombinations of such method, including the particular analytical method disclosed herein for determining car" bon residue. 7

The exact nature of the components of the oil that are responsible for the carbon residue of the oil is not definitely known. Nor is it definitely known Whether the components that are responsible for the characteristic absorption of light upon which the analytical method of the present invention is based are the same as or different from the components that actually are responsible for carbon residue. However, it is clear from experiments carried out upon a large number of hydrocarbon oil samples having a 'wide range of carbon residues that the components that exhibit a characteristic absorption for light of the wavelengths disclosed herein are at least present in an amount that varies with the amount of the components that are responsible for the carbon residue of the oil, and more than likely, the components are substantially the same.

It is to be emphasized that the chemical compounds in the oil that are responsible for the characteristic light absorption utilized in the analyticalmethod of the present invention comprise a complex mixture of materials of chemical structure that is not precisely known, the individual components of which apparently differ but slightly from member to member over a relatively broad range of compounds. This fact is important as it renders the measurement of the amount of such compounds by spectrophotometry completely different from the spectrophotometric methods heretofore used for quantitative measurement of other materials, for example, single compounds or other mixtures. Thus, whereas pure compounds and many mixtures are usually found to exhibit one or more characteristic absorption peaks that are typical of such compound or compounds, the complex ixtures of compounds with which the present invention is concerned exhibit no such characteristic peaks. Rather, because of the presence in the mixture of a relatively large number of compounds that differ but slightly from member to member and that therefore absorb adjacent wavelength of light, the absorption curve for the compounds in question and for the compounds adjacent thereto is ordinarily a more or less smooth curve, frequently approaching a straight line. Thus, the eye is normally unable to distinguish any significant dilference between the portion of the absorption curve that is responsible for the characteristic absorption utilized in the analytical method of the present invention and the portion of the absorption curve that bears no relation to the car bon residue of the oil. The present invention is based on the discovery of the fact that specific portions of the light absorption curve for a hydrocarbon oil that are not characterized by any distinguishing shape, furnish a quantitative index of components in the oil that are indicative of the carbon residue content of the oil.

eferring now briefly to the drawings, in FIGURE 1 there is shown a calibration curve plotting absorption signal intensity for light having a wavelength of about 398 millimicrons against Ramsbottom carbon residue content for a variety of petroleum gas oils. FIGURE 2 is a calibration curve plotting Conradson carbon residue content against the integrated absorptivity per liter per gram centimeter for light of a plurality of wavelengths representative of the 350 to 400 millimicron band for a number of petroleum gas oils. FIGURE 3 is a simplified flow diagram of a vacuum distillation tower of the kind used in the preparation of a gas oil charge stock for a fluid catalytic cracking unit, wherein certain conditions of the distillation are controlled in response to changes in light absorption, that is, changes in carbon residue, of an oil stream subjected to analysis in accordance with this invention.

In accordance with the analytical method disclosed herein, there is first obtained a sample of a hydrocarbon oil having unknown carbon residue characteristics. Sampling may be carried out in any convenient Way. Thus, in the case of a process stream, the oil may be sampled intermittently and the thus-obtained samples placed in the test cell of a spectrophotometer and the desired measurement obtained on a batch-wise basis. Alternatively, the hydrocarbon oil stream can be sampled continuously and a continuous stream of the oil can be circulated through the test cell. In either case, when the oil is quite dark it is desirable to dilute the oil prior to analysis with a solvent that will not itself absorb a significant proportion of the radiation of the wavelength employed in the analysis, that is, that Will not interfere with the selective absorption by the components in the oil that are indicative of carbon residue, so that a measurable amount of radiation will be transmitted to form a detectable electric-al output signal. Clear naphtha has been used successfully', but other solvents that do not interfere with the selective absorption of radiation by the oil being tested can be used. Specific examples of other solvents include mononuclear aromatic solvents such as benzene and toluene and paraffinic solvents such as kerosene. The proportion in which the solvent is employed will depend upon the absorptivity of the oil and the length of the radiation path through the test cell of the spectrophotometer. Good results have been obtained, for example, with dilutions of 9 to 24 parts naphtha solvent per part of a heavy :gas oil that prior to dilution is opaque to radiation of the Wavelength employed in the analysis, using test cell path lengths varying from 0.5 millimeter to 2 centimeters in length, but other dilutions and other test cell path lengths can be used. In the event that continuous sampling, as well as dilution, are desired, these objectives can be achieved by the use of conventional proportioning pump means designed to pump and blend two fluids in a predetermined ratio. Examples of suitable proportioning pumps are shown in United States Patent No. 2,873,889 to Mori, but other functionally equivalent pumps can be used.

The temperature at which the oil is subjected to spectrometry is not critical in principle, except of course that the temperature should be suificiently low that portions of the oil will not be olatilized, although not so low that the oil will solidify. However, the limitations of commercially available spectrophotometric apparatus are presently such that best results are obtained when the oil is at a temperature not greater than about 200 F. Excellent results have been obtained with diluted oil samples at ambient atmospheric temperature, but of course, good results also can be obtained at other temperatures. It will be understood that the oil streams in many processing operations for which the carbon residue is to be determined will be at temperatures greater than 200 F. and accordingly, samples obtained from tlese streams for analysis in accordance with the present invention must be cooled before subjecting them to analysis in accordance with the herein-disclosed method. How ever, this requirement poses no real problem, and cooling of the stream to the desired test temperature can be effected rapidly and continuously, when desired, by conventional heat exchange devices, condensers, or the like.

The radiation to which the oil is subjected should be of a wavelength that is selectively absorbed by the components that are indicative of carbon residue, that is to say, of a Wavelength that will be absorbed by the aforesaid components Without at the same time being appreciably absorbed by other components of the oil that are not indicative of carbon residue. As indicated, excellent results have been obtained by the use of light having a plurality of wavelengths representative of the band of about 350 to 400 millimicrons. Light of somewhat higher and lower wavelength can be used, but the accuracy of the correlation between carbon residue and light absorption decreases rather rapidly outside the Wavelength band indicated. Preferably, the Wavelength of the light is not appreciably less than 350 millimicrons as inclusion of light of such shorter wavelengths appears unduly to distort the carbon residue values obtained by the present method. Inasmuch as the light absorption for the matflElllS in question varies less greatly with changes in Wavelength near the 400 millimicron limit of the Wavelength band, a somewhat greater departure from the above-indicated upper limit, for example, up to 405 millimicrons and possibly even as high as 425 millimicrons, can be tolerated. When using light in the 350 to 400 mill-imicron wavelength band, for the purpose disclosed, excellent results can be obtained by scanning the test oil with light having a plurality of wavelengths that are representative of the wavelength band indicated, plotting absorption signal intensity for each such wavelength, or a function thereof, such as absorbance or absorptivity that is also related to the concentration in the oil of the components that are indicative of carbon residue, as a function of the wavelength, and integrating the thus-obtained curve to determine the area thereunder. Absorptivity is defined as absorbance per unit concentration of the material subjected to analysis in grams per liter, per unit test cell path length in centimeters. Absorbance is defined as the logarithm to the base of the reciprocal of the transmittance. The area under the absorptivity curve the 350 to 400 millimicron wavelength band has been found to be related quantitatively to the carbon residue of the oil undergoing analysis. With some oils slight peaks of varying intensity sometimes have been observed in the absorption curve in the 350 to 400 millimicron Wavelength band that are not attributable to components indicative of carbon residue. However, by correlating carbon residue with a function of the absorption signal intensity over the entire 359 to 400 millimicrcn wavelength band, inaccuracies resulting from absorption of radiation by substances that are not indicative of carbon residue are minimized.

Although good accuracy has been obtained from the correlation of carbon residue and integrated abscrptivity, that is, a function of the absorption signal intensity reflecting absorption over the entire Wavelength band of 350 to 400 niillimicrons, this particular correlation is not absolut ly essential to the invention, and good results also can be obtained by cor elation between carbon residue and the integrated absorbance over the Wavelength band indicated, or the absorption signal intensity at a single wavelength of lig t, or other functions of such signal intensity that are related to the concentration in the oil of components that are indicative of carbon residue con tent, for example, absorptivity. Especially good results are obtainable from the absorption signal intensity at a single wavelength of about 400 millimicrons, but some deviation from this wavelength can be tolerated. For direct correlation of carbon residue at a single absorption signal intensity, we prefer to use light having a wavelength of about 393 to 405 millirnicrons, with light having a wavelength of about 398 to 400 millimicrons being especially useful.

For purposes of hydrocarbon oil refining process control, it is not absolutely necessary that the carbon residue content for the particular oil undergoing analysis actually be determined. Once the desired efiiciency in the processing operation in question has been obtained, whatever the carbon residue of the stream undergoing analysis at that time may be, it is sufiicient thereafter merely to effect control of the desired processing operation. according to whether the absorption signal intensity or the selected function thereof increases or decreases. As indicated the amount of the process adjustment will preferably be governed either by the integrated absorptivity using light having a wavelength band of 350 to 400 millimicrons, or directly with the absorption signal intensity using light having a wavelength of about 400 millimicrons.

Although it is not necesasry for purposes of process control to determine the actual numerical carbon residue value of the oil undergoing analysis, it is sometimes desirable to do so as a partial indication of the quality of such oil. In these instances the actual numerical carbon residue content of the oil can easily be obtained by the use of a preestablished correlation between carbon residue content and absorption signal intensity or a function thereof that is also related to the concentration in the oil of components that are indicative of carbon residue. This correlation can conveniently take the form of a calibration curve, table or equivalents thereof.

in determining the absorption of light by the oil samples for purposes of this invention, conventional spec trophotometric apparatus can be used. Thus, when light having a single wavelength is to be used, either conventional filter-type or dispersion-type spectrophotometers can be used. When absorption of light having a plurality of wavelengths over a Wavelength band is to be determined, the dispersion-type spectrophotometer is usually most convenient, but the filter-type spectrophotometer can be used provided that a filter capable of transmitting light of the desired wavelength band is available. In the latter instance a single absorption signal intensity or a function thereof representing the combined absorption signal intensities over the entire wavelength band will be obtained rather than a curve of absorption signal intensities or functions thereof over the wavelength band.

The analytical process of the present invention is useful as such merely to determine carbon residue, but it finds especial utility as a means of process control in hydrocarbon oil refining operations. For example, the analytical process described herein can be used to control the cut-point in the vacuum distillation of topped crude oil to obtain a heavy gas oil component of the charge to a catalytic cracking unit. As the boiling point of the highest boiling side stream from the vacuum distillation tower increases, so also does the carbon residue content of that stream and the carbon residue content of the total catalytic cracking unit feed of which it forms a part. As the carbon residue of the feed to the catalytic cracking unit increases, so also does the amount of coke deposited upon the catalyst, usually on a percent for percent basis, the coke on catalyst in this comparison being measured as percent coke based on the wei ht of the fuel. An increase in the amount of coke that is deposited upon the cracking catalyst that cannot be compensated for by reserve regeneration capacity, as is the usual case, is accompanied by a reduction in cracking catalyst activity. In accordance with the present invention, as the carbon residue content of the heavy gas oil and/or catalytic racking unit feed stock increase to an undesirable level, the absorption signal intensity of such feed stock will also increase and such increase can be caused by means of conventional recorder-controller equipment to effect a change in a processing available to correct or compensate for the change in carbon residue content of the oil. For example, in one embodiment the fuel flow to the heater for the vacuum tower feed can be reduced so as to reduce the cut-point, i.e., reduce the boi ing point of the highest boiling side stream of the vacuum distillation tower. Alternatively, the severity of the cracking conditions can arenas? be reduced, or, when some reserve regeneration capacity is available, an increase in the absorption Slg .1 intensity can be used to increase the regeneration air, regeneration time, and/ or regeneration temperature during the regeneration of the cracking catalyst, so as to burn off the additional amount of colic deposited on the catalyst by the feed. Conversely, as the carbon residue or" the hydrocarbon oil stream undergonng analysis decreases, the cut-point of the highest boiling vacuum tower stream can be raised, or alternatively, the severity of t.e cracking conditions can be increased, or the severity of the regenerating conditions can be reduced, so as .0 maintain the optimum conversion depth. Tl e analytical method of the present invention can also be used in a similar manner to control other hydrocarbon oil processing operations. For example, it can be used to control the cut-point in the vacuum distillation of lubricating oil stocks, or to control the exhaustiveness of solvent treatment in the solvent refining of lubricating oil stocks.

The operability of the analytical method of the present invention has been demonstrated by calculat' 2g carbon residue values of a number of gas oils in accordance with the present method and comparing the values obtained with carbon residue values obtained by conventional laboratory methods. In carrying out some of these experiments, an Analytical Systems inc. filter-type spectrophotometer Was employed utilizing a Baird-Atomic Interference Filter with a band pass at 398 millimicons and cut-01f transmission at 393 and 405 millirnicrons. A 4-97 Corning glass color filter was used in coniunction with the Baird filter to block out transmission of the Baird filter about 690 miliimicrons. In these experiments the output signal from the spectrophotometer was set to vary between and millivolts for 100 to 0 percent transmission.

In accordance with the experimental procedure followed, one milliliter of each gas oil to be analyzed was diluted to milliliters with a heavy Kuwait naphtha so that some light of the wavelength selected would be transmitted through the otherwise opaque gas oil samples. Each diluted gas oil sample was then placed in the absorption cell of the spectrophotometer.

1 no test cell in this instance had a path length of 0.5 millimeter, and the absorption signal intensity scale reading was observed on the recording scale of a Speedomax (Leeds & Northrop, Inc.) 0-5 millivolt recorder.

The absorption signal density at a wavelength of 398 millimicrons in the scale units of the recording instrument was obtained for 39 different gas oils having Ramsbottorn carbon residues from 0.22 to 1.39 percent. A calibration curve was established by plotting the absorption signal intensity in recorder scale units as a function of Ramsbottorn carbon residues for the oils undergoing analysis. The Ramsb ttom carbon residues employed in the preparation of the calibration curve were averages of three or more laboratory determinations. A straight 1 13 was drawn through the thus-plotted points, using the so-called least squares method. A plot of the calibration curve obtained in this manner is shown in FlGURE 1.

To determine the accuracy or" he method described herein, the absorption signal intensity in recorder scale units for each of the 39 gas oils tested was applied to the calibration curve obtained as described above to obtain a calculato carbon residue value. The difference between the thus-obtained calculated carbon residue value and the average Ramsbottom carbon residue value for the same oil was then determined and the percent deviation from the laboratory value was calculated. The recorder scale units, which are an index of percent light absorption, obtained for each gas oil sample in the abovedescribed analysis, the corresponding Ramsbottom carbon residue values, the calculated carbon residue values, the difference between calculated and Rarnsbottom carbon residues, and tie percent deviation between the calculated the fol- T able A Scale Ramsbtm. Gale. A carbon Percent Sample No. units carbon carbon residue deviation residue residue 27.5 0.47 0. 4G 0. 01 2.1 35. 0 0. 60 0. 02 +0. 02 3. 3 26. 5 0. 45 0.44 -0. 01 2. 2 22.0 0.41 0.36 +0. 05 12. 2 27. 5 0. 45 0. 47 +0.02 4. 4 20.0 0.35 0.32 -0.03 8. 6 80.0 0.65 0.6i -0.01 1.5 21. 0 0.32 0. 34 +0.02 0.3 20. 0 0. 37 0. 44 +0.07 18. 9 21.0 0.35 0. 34 -0.01 2.9 22.0 0.37 0. 36 0. 01 2. 7 33. 0 0.47 0. 58 +0.11 23. 4 26. 5 0. 38 0. 45 +0. 07 18. 4 26. 5 0. 44 0. 4.5 +0.01 2. 3 21. 0 0.27 0.34 +0.07 25. 9 15.0 0.22 0.22 0.00 0.0 20. 5 0.26 0.32 +0. 06 23.1 14.0 0.23 0.19 0. 04 17. 4 21. 5 0.30 0.3i +0.04 l3.3 14. 0 0.25 0. 20 t). 05 20.0 40. 0 0. 64 0.71 +0.07 10. 0 23.5 0.43 0. 38 0. 05 11.6 32. 5 0. 59 0. 56 0. 03 5.1 19.0 0.35 0.30 0. 05 14. 3 35. 0 0.57 0. 01 +0. 04 7.0 20. 5 0.33 0.32 0. 01 3.0 31.0 0.52 0. 53 +0.01 1. 9 22. 5 0.37 0. 36 0. 01 2. 7 31. 5 0. 44 O. 54 +0.10 22. 7 22.0 0. 36 0.35 0. 01 2. 8 31.5 0. 46 0. 54 +0.08 17.4 24.0 0. 1 0.39 +0.02 4.9 57.0 1. 03 1. 06 +0. 03 2. 9 50. 0 0.82 0. 91 +0.09 11.0 69. 0 1. 39 1. 30 -0. 09 6. 5 40. 0 0. 76 0.71 0. 05 6. 6 47. 0 1.19 0.85 +0. 34 2s. 6 44. 5 0.71 0.80 +0. 09 12. 7 77.0 1. 26 1. 45 +0. 19 15. 1

From the foregoing table it will be seen that the standard, or average, deviation of the calculated carbon residue values from the laboratory carbon residue value is about 13 percent. This compares favorably with the standard deviation between successive determinations by laboratory methods on the same oil.

To demonstrate the correlation between carbon residue content and integrated absorptivity over the 350 to 400 millimicron wavelength band, integrated absorptivity values in the wavelength band indicated were obtained for 20 different gas oil samples having Conradson carbon residues of 0.21 to 1.23 percent. A calibration curve was obtained from these data by plotting the integrated absorptivity as a function of Conradson carbon residues, and by drawing a straight line through the thus-plotted points, again using the least squares method. The carbon residue contents of the different gas oils were then calculated from the integrated absorptivities for these oils using the thus-prepared calibration curve. The calculated carbon residues were then compared with the laboratory carbon residues as described above. The calibration curve for this series of experiments is demonstrated in FEGURE 2. The term absorptivity area employed in FIGURE 2 corresponds to the integrated absorptivity and refers to the area under the curve obtained by plotting absorptivity against wavelength.

In the series of experiments referred to in the preceding paragraph, a Cary Model 11 dispersion-type spectrophotometer, manufactured by the Applied Physics Corporation was employed, using a one-centimeter pathlen-gth quartz sample tube. Unlike the Analyti al Sys tems Inc. spectrophotometer described hereinabove, which produces light of the desired wavelength by the use of a filter capable of transmitting only light of the desired wavelength, the dispersion-type spectrophotometer produces light of the desired wavelength by dispersing individual wavelengths of light from a source of mixed wavelengths by means of a suitable prism. The prism is rotated to focus light of the desired wavelength upon a mirror which reflects the desired wavelength of light through suitable monochomating means to eliminate radiation of undesired wavelength and thence through the test samples to a photomultiplier for detecting the transmitted light and converting it to an electrical output signal. Since the wavelength of the light passed through the test sample can be easily varied by rotating the dispersing prism in appropriate fashion, the dispersion-type spectrophotometer is well-suited both for determining absorption by the test sample of light of a number of wavelengths that are representative of a broad wavelength band, as well as of light having a single Wavelength.

The Conradson carbon residue values, calculated carbon residue values, and the percent deviation of the calculated carbon residue values from the Conradson carbon residue values are shown in the following Table B:

Table B Con- 350-400 Milliradson microns Sample Description Carbon Residue (0.1%.) Gale. A per- O.R. cent Mid-Continent Gas Oil 0. 53 0. 52 -2 West Texas Gas Oil 0.68 0.87 +28 Heavy Overhead Disti1late.. 0.50 0.49 -2 Blend of Three Kuwait Gas Oi1s- 0.75 0.07 -11 (3.0. tr. Normal Vac. Reduction of Gasoline-Free Reduced Mara- W. Voncz. Crude 0.52 0.40 12 West. Venez. Virgin Gas 1.06 0.88 -17 Composite Typical Chg. Stock to 0. 46 0. 46

Fluid Unit; Still. Virgin Kuwait Gas Oil... 0. 59 0. 53 Fluid Unit Chg. Stock. 0. 64 0.55 14 Virgin Mld-Cont. Gas 01 0.31 0.21 32 Virgin Kuwait Gas Oil... 0. 54 0.48 -11 Virgin West Texas Gas Oil 0.29 0.32 +10 STliouisiana Overhead (721-053" 0. 53 0. 54 +2 4 S.}Louisiana Overhead (721-1,000 0.68 0.63 -7 I). Cepta (Eocene) Ovhd. (HO-950 1.01 1. 08 +7 1. CeIuLa (Eocene) Ovhd. (730-972" 0.91 1.07 +18 oi 'ti (Eocene) Ovhd. res-1,000" 1.20 1.23 +3 18 Ovhd. from Ragusa Crude 0.35 0.47 +34 19 (lo 0. 51 0. 58 +14 20 .d0 0.69 0.71 +3 From the foregoing table it will be seen that the standard or average percent deviation of the carbon residue values obtained by the herein-disclosed spectrophotometric analytical method from the laboratory carbon residues was about 16 percent. Again, this deviation compares favorably with the percent deviation between repeated laboratory determinations on the same gas oil. A still smaller percent deviation can be obtained by correlating the absorptivity area as obtained above with Ramsbottom carbon residues instead of Conradson carbon residues, as the dispersion of the results obtained by the former laboratory method is relatively smaller.

in a working embodiment, with particular reference to FIGURE 3, 346.5 barrels per hour of the distillation bottoms from an atmospheric distillation tower, not shown, are passed through line 2 and blended with 13.8 barrels per hour of recycle oil from line '48, obtained from the bottom tray 1.3 of the fractionation Zone of vacuum distillation tower l6, and the mixture of atmospheric tower bottoms and vacuum recycle oil is changed through line 4 to a vaporizing heater 6 whose burner is supplied with fuel from line 3. The heated feed is then passed into line 12 where it is mixed with steam from line 14 to assist in vaporizing the heavier components of the oil, and the over-all feed is then passed to the flash zone 14 of the vacuum tower 16 at a temperature of 757 F. in the dash zone 14, which is maintained at an absolute pressure of about 74 millimeters Hg, the vaporizable components of the feed stock are flashed off from the unvaporized bottoms and are passed upwardly into the fractionation zone. As the vapors pass upwardly, they are washed with liquid retained in

fractionation trays

13, 20, 22, 2d, 2d, and 28, the heavier portion of the oil vapors being condensed in the liquid in the trays contacted thereby. Two oil side streams, in the respective amounts of 107.1 barrels per hour and 97.5 barrels per hour are withdrawn by way of

lines

33 and 42 and 32 and 36, respectively, and passed into line 50 and thence into line 54 where the combined side streams are blended with 341.3 barrels per hour of a light gas oil stream obtained from line and from the atmospheric distillation tower, not shown. Reflux streams from the respective side streams 33 and 32 are returned to the vacuum tower 16 by way of

lines

40 and 34 at rates of 489 and 280 barrels per hour, respectively. A side stream of oil is also removed from fractionation tray 18 by way of line 44. A portion of this oil is returned to the space above tray 13 by way of line as, and the remainder is passed through inc 4% as recycle oil.

During the operation of the unit as described above, a sample stream is Withdrawn from line 5 through line 56 and cooled to F. in cooler 53. The cooled oil is then passed by way of line 60 through one side of a gear-type proportioning pump 62 and thence into line as where each volume of cooled oil is blended with 24 parts by volume of naphtha drawn from line by the other side of the proportioning pump e2. The cooled and diluted gas oil sample is then continuously passed through the sample tube of the carbon residue analyzer of spectrophotometer 68. In this instance the carbon residue analyzer is a spectrophotometer of the nondispersion-type employing an interference filter with a band pass at 398 millimicrons and cut-oil transmission at 393 and 405 millimicrons, employed in coniunction with a Corning 4-97 glass color filtcr to block out transmission of the first described filter above 690 millimicrons. In this embodiment the output signal from the analyzer is set for 0 to 5 millivolts corresponding to 100 to 0 percent transmission. The light transmitted through the sample tube of the analyzer 68 is electrically converted to an electrical signal whose magnitude is proportional to the amount of light absorbed by the diluted 'gas oil in the sample tube. After passing through the analyzer the cooled diluted gas oil sample can be passed to waste oil storage through line 70, or alternatively, since it is relatively small in amount, the diluted gas oil can be returned to line 54 through a conduit not shown.

The signal from analyzer as is passed to a recordercontroller '72 of conventional construction where the magnitude of the signal is recorded quantitatively on the recording scale. In the present embodiment the unit is set to operate with a total catalytic cracking feed Ramsbottom carbon residue of about 0.50, which corresponds to a yield of about 1.07 colre on cracking catalyst based on the weight of the oil at a conversion depth of 50 percent, a reactor temperature of 925 and at a carbonon-regenerated catalyst level of 0.2 weight percent. :In the embodiment described, this operation corresponds to a scale reading on the recorder of approximately 30 scale units. Under these conditions, the quantity of fuel passed through line it into vacuum tower heater 6 is controlled at the desired rate by means of pneumatic diaphragn motor valve 10 which in turn is controlled by the supply of instrument air from line 74, which in turn is governed by the recorder-controller '72, :hich is of the recording potentiometer-pneumatic controller type.

As the carbon residue of the total catalytic cracking unit feed through line 54 varies upwardly or downwardly a proportionate and compensatory adjustment in the amount of heater fuel passed through valve 10 is brought about.

Under an alternative method of operation, instrument air from recorder-controller "i2 and line 74 is directed through dashed line '76 from which it is caused to control pneumatic diaphragm motor valve 73. As valve '78 is unged to a more closed position, the amount of recycled oil through line to the upper surface of tray 22 is increased. As the amount of liquid returned to tray 22 increases, so also the amount of liquid flowing downwardly to the trays lower in the column increases, as a consequence of which the amount of recycled oil through line is increased. As a result of this operation the proportion in the total catalytic cracking charge stock represented by the highest boiling side stream of the vacuum tower is is reduced, whereby the carbon residue of the total catalytic cracking charge stock is also reduced. Conversely, urging of valve to a more open position will increase the proportion in the total catalytic cracking feed of the highest boiling side stream of the vacuum tower, whereby the carbon residue of the catalytic cracking feed will be increase The analytical method of the present invention is adaptable to the hydrocarbon oil refining process control by virtue of the fact that it can be carried out rapidly and continuously Without destruction of the sample undergoing analysis. Although the method is readily adaptable to automatic control, it is also useful per so as a laboratory analytical method in place of presently used laboratory methods for determining carbon residue values.

The expression hydrocarbon oil is employed herein in its usual sense to define oils that are composed chiefly of components containing essentially carbon and hydrogen, of which petroleum oils and fractions derived therefrom are typical samples. The expression is not intended to exclude oils the contain minor proportions of elements other than carbon or hydrogen such as sulfur, oxygen, nitrogen, etc., as is the usual case with petrolum oils and fractions thereof.

By the expression hydrocarbon oil refining process and the like is meant any process for upgrading hydrocarbon oils or fractions thereof, including processes involving distillation, blending, solvent treating, and chemical and catalytic treating processes.

The invention is not limited to the embodiments shown and described herein. Many other modifications will occur to those skilled in the art, and such modifications can be resorted to without departing from the spirit or scope of this invention. Accordingly, only such limitations should be imposed as are indicated in the appended claims.

We claim:

1. A method for controlling a hydrocarbon oil refining process in response to changes in the carbon residue content of an oil stream of said process, comprising subjecting a plurality of samples talren at difierent times from said oil stream to spectrometric analysis by passing light therethrough of a wavelength that is selectively absorbable by components of the oil that are indicative of the carbon residue content of the oil, in a proportion related to the quantity of said components in the oil, detecting the intensity of the light transmitted through the oil samples, and converting the transmitted light to at least one ouput signal whose intensity is related to the quantity of said components in the oil samples, and modifying a variable in said hydrocarbon oil refining process in response to changes in at least one of (a) the absorption signal intensity, and (b) a function thereof that is also related to the quantity of said components in the oil samples.

2. A method for controlling a hydrocarbon oil refining process in response to changes in the carbon residue content of an oil stream of said process, comprising subjecting a plurality of samples taken at ditlerent times from said oil stream to spectrometric analysis by passing light therethrough having a plurality of Wavelengths that are representative of the Wavelength band of about 350 to 400 millimicrons and that is selectively absorbable by components of the oil that are indicative of the carbon residue content of the oil, in a proportion related to the quantity of said components in the oil, detecting the intensity of the light of such wavelengths transmitted by the oil samples, and converting the transmitted light to output signals whose intensity is related to the quantity of said components in the oil samples, and modifying a variable in said hydrocarbon oil refining process in response to changes in at least one of (a) the absorption signal intensities for the respective oil samples, and (b) a function thereof that is also related to the quantity of said components in the oil samples.

3. A method for controlling a hydrocarbon oil relinin g process in response to changes in the carbon residue content of an oil stream of said process, comprising subjecting a plurality of samples taken at difierent times from said oil stream to spectrometric analysis by passing light trerethrough having a plurality of Wavelengths that are respresentative of the Wavelength band of about 350 to 400 millimicrons and that is selectively absorbable by components of the oil that are indicative of the carbon residue content of the oil, in a proportion related to the quantity of said components in the oil, detecting the intensity of the light of such wavelengths transmitted by the oil samples, and converting the transmitted light to output signals Whose intensity is related to the quantity of said components in the oil samples, plotting one of (a) the absorption signal intensity, and (b) a function thereof that is also related to the quantity of said components in the oil samples, for said wavelengths of light to obtain a curve indicative of the light absorption in the wave length band indicated, determining the area under said curve, and modifying a variable in said hydrocarbon oil refining process in response to changes in said area for said oil samples.

4. A method for controlling a hydrocarbon oil refining process in response to changes in the carbon residue content of an oil stream of said process, comprising subjecting a plurality of samples taken at diiierent times from said oil stream to spectrometric analysis by passing light therethrough having a Wavelength of about 400 millimicrons and that is selectively absorbable by components of the oil that are indicative of the carbon residue content of the oil, in a proportion related to the amount of said components in the oil, detecting the intensity of the light transmitted by the oil samples, and converting the transmitted light to output signals Whose intensity is related to the quantity of said components in the oil samples, and modifying a variable in said hydrocarbon oil refining process in response to changes in at least one of (a) the absorption signal intensity and (b) a function thereof that is also related to the quantity of said components in the oil samples.

5. A method for controlling a hydrocarbon oil refining process in response to changes in the carbon residue content of a distillate oil stream of said process, comprising subjecting a plurality of samples taken at different times from said distillate oil stream to spectrometric analysis by passing light therethrough having a wavelength that is selectively abso-rbablc by components of the oil that are indicative of the carbon residue content of the oil, in a proprotion related to the quantity of said components in the oil, detecting the intensity of the light transmitted by the oil samples and converting the transmitted light to output signals whose intensity is related to the quantity of said components in the oil samples, and modifying the distillation conditions under which said distillate hydrocarbon oil stream is produced in response to changes in at least one of (a) the absorption signal intensity, and (b) a function thereof that is also related to the quantity of said components in the oil samples.

6. A method for quantitatively analyzing a hydrocarbon oil for carbon residue content comprising subjecting a sample of the hydrocarbon oil to spectrometric analysis by passing light therethrough having a wavelength that is selectively absorbable by components of the oil that are indicative of the carbon residue content of the oil, in a proportion related to the quantity of said components in the oil, detecting the intensity of the light transmitted by the oil and converting the transmitted light to at least one output signal Whose intensity is related to the quantity of said components in the oil, and determining the carbon residue content that corresponds to one of (a) the absorption signal intensity, and (b) a function thereof that is also related to the quantity of said components in the oil, as shown by a pro-established correlation therebetween.

7. A method for quantitatively analyzing a hydrocarbon oil for carbon residue content comprising subjecting a sample of the hydrocarbon oil to spectrometric analysis by passing light therethrough having a plurality of wavelengths representative of the Wavelength band of about 350 to 400 millimicrons and that is selectively absorbable by components of the oil that are indicative of the carbon residue content of the oil, in a proportion related to the quantity of said components in the oil, detecting the intensity of the light of such wavelengths transmitted by the oil and converting the transmited light to output signals whose intensity is related to the quantity of said components in the oil, and determining the carbon residue content of the oil that corresponds to one of (a) the absorption signal intensity, and (b) a function thereof that is also related to the quantity of said components in the oil, as shown by a pre-established correlation therebetween.

8. A method for quantitatively analyzing a hydrocarbon oil for carbon residue content, comprising subiecting a sample of the hydrocarbon oil to spectrometric analysis by passing light therethrough having a plurality of wavelengths representative of the wavelength band of about 350 to 400 millimicrons and that is selectively absorbable by components of the oil that are indicative of the carbon residue content of the oil, in a proportion related to the quantity of said components in the oil, detecting the intensity of the light of such wavelengths transmitted by the oil, and converting the transmitted light to output signals whose intensity is related to the quantity of said components in the samples, plotting one of (a) absorption signal intensity, and (b) a function thereof that is also related to the quantity of said components in the oil, for said wavelengths of light to obtain a curve indicative of light absorption in the indicated Wavelength band, determining the area under said curve, and determining the carbon residue content that corresponds to the thusdetermined area as shown by pre-established correlation between carbon residue and the area under the absorption curve.

9. A method for quantitatively analyzing a hydrocarbon oil for carbon residue content comprising subjecting a sample of the hydrocarbon oil to spectrometric analysis by passing light thercthrough having a wave length of about 400 millimicrons and that is selectively absorbable by components of the oil that are indicative of the carbon residue content of the oil, in a proportion related to the amount of said components in the oil, detecting the intensity of the light transmitted by the oil, and converting the transmitted light to an output signal whose intensity is related to the quantity of said components in the oil, and determining the carbon residue content of the oil that corresponds to one of (a) the absorption signal intensity, and (b) a function of absorption signal intensity that is also related to the quantity of said components in the oil as shown by a preestablished correlation therebet-ween.

10. An on-line analytical apparatus for quantitatively analyzing a hydrocarbon oil process stream for carbon residue content, comprising hydrocarbon oil refining apparatus provided with conduit means for conducting a hydrocarbon oil stream in said refining apparatus, sampling means associated with said conduit means for withdrawing hydrocarbon oil samples from the oil stream transported thereby, speotrophotometric means provided with an absorption cell connected to said sampling means so as to permit flow of said oil samples into said absorption cell, said spectrophotometric means also being provided with a source of light having a wavelength that is selectively absorbable by components of the oil that are indicative of the carbon residue content of the oil, and that is selectively absorbable by such components in a proportion related to the quantity of said components in the oil, said spectrophotometric means also including detecting means for sensing the intensity of the light of said wavelength transmitted by the oil, and means for converting the transmitted light to an output signal whose intensity is related to the intensity of the transmitted light and thus to the quantity of said components in the oil, and recording means for indicating at least one of (a) the absorption signal intensity, and (b) a function thereof that is also related to the quantity of said components in the oil.

11. The apparatus of claim 10 where said source of light has a plurality of wavelengths representative of the wavelength band of about 350 to 400 millimicrons.

12. The apparatus of claim 10 where said source of light has a wavelength of about 400 millimicrons.

References Cited in the file of this patent UNITED STATES PATENTS 2,047,985 Weir July 21, 1936 2,753,292 Porter et al. July 3, 1956 2,766,265 Goldsmith et a1 Oct. 9, 1956 2,906,798 Starnes et al. Sept. 29, 1959 2,910,523 Montgomery et al. Oct. 27, 1959 2,994,646 Kleiss Aug. 1, 1961

Claims

1. A METHOD FOR CONTROLING A HYDROCARBON OIL REFINING PROCESS IN RESPONSE TO CHANGES IN THE CARBON RESIDUE CONTENT OF AN OIL STREAM OF SAID PROCESS, COMPRISING SUBJECTING A PLURALITY OF SAMPLES TAKEN AT DIFFERENT TIMES FROM SAID OIL STREAM TO SPECTROMETIC ANALYSIS BY PASSING LIGHT THERETHROUGH OF A WAVELENGTH THAT IS SELECTIVELY ABSORBABLE BY COMPONENTS OF THE OIL THAT ARE INDICATIVE OF THE CARBON RESIDUE CONTENT OF THE OIL, IN A PROPORTION RELATED TO THE QUANTITY OF SAID COMPONENTS IN THE OIL, DETECTING THE INTENSITY OF THE LIGHT TRANSMITTED THROUGH THE OIL SAMPLES, AND CONVERTING THE TRANSMITTED LIGHT TO AT LEAST ONE OUPUT SIGNAL WHOSE INTENSITY IS RELATED TO THE QUANTITY OF SAID COMPONENTS IN THE OIL SAMPLES, AND MODIFIYING A VARIABLE IN SAID HYDROCARBON OIL REFINING PROCESS IN RESPONSE TO CHANGES IN AT LEAST ONE OF (A) THE ABSORPTION SIGNAL INTENSITY, AND (B) A FUNCTION THEREOF THAT IS ALSO RELATED TO THE QUANTITY OF SAID COMPONENTS IN THE OIL SAMPLES.

10. AN ON-LINE ANALYTICAL APPARATUS FOR QUANTITATIVELY ANALYZING A HYDROCARBON OIL PROCES STREAM FOR CARBON RESIDUE CONTENT, COMPRISING HYDROCARBON OIL REFINING APPARATUS PROVIDE WITH CONDUIT MEANS FORM CONDUCTING A HYDROCARBON OIL STREAM IN SAID REFINING APPARATUS, SAMPLING MEANS ASSOCIATED WITH SAID CONDUIT MEANS FOR WITHDRAWING HYDROCARBON OIL SAMPLES FROM THE OIL STREAM TRANSPORTED THEREBY, SPECTROPHOTOMETIC MEANS PROVIDED WITH AN ABSORPTION CELL CONNECTED TO SAID SAMPLING MEANS SO AS TO PERMIT FLOW OF SAID OIL SAMPLES INTO SAID ABSORPTION CELL, SAID SPECTROPHOTOMETRIC MEANS ALSO BEING PROVIDED WITH A SOURCE OF LIGHT HAVING A WAVELENGTH THAT IS SELECTIVELY ABSORBABLE BY COMPONENTS OF THE OIL THAT ARE INDICATIVE OF THE CARBON RESIDUE CONTENT OF THE OIL, AND THAT IS SELECTIVELY ABSORBABLE BY SUCH COMPONENTS IN A PROPORTION RELATED TO THE QUANTITY OF SAID COMPONENTS IN THE OIL, SAID SPECTROPHOTOMETRIC MEANS ALSO INCLUDING DETECTING MEANS FOR SENSING THE INTENSITY OF THE LIGHT OF SAID WAVELENGTH TRANSMITTED BY THE OIL, AND MEANS FOR CONVERTING THE TRANSMITTED LIGHT TO AN OUTPUT SIGNAL WHOSE INTENSITY IS RELATED TO THE INTENSITY OF THE TRANSMITTED LIGHT AND THUS TO THE QUANTITY OF SAID COMPONENTS IN THE OIL, AND RECORDING MEANS FOR INDICATING AT LEAST ONE OF (A) THE ABSORPTION SIGNAL INTENSITY, AND (B) A FUNCTION THEREOF THAT IS ALSO RELATED TO THE QUANTITY OF SID COMPONENTS IN THE OIL.