Method for optimizing control of a production unit

Abstract

The octane number of a product is directly determined from its near infrared (NIR) absorption spectrum in the wave number spectral range of from 6667 to 3840 centimeter .sup.-1. A number (n) of frequencies is selected within this range, and the (n) absorbance values are correlated with the octane number by using a multivariate regression analysis.

Description

The invention relates to a method for the determination of the octane number of petroleum products using near infrared (NIR) spectroscopy.

There are several octane numbers, including the motor octane number and the research octane number, defined in particular by the French Standard NF M07-026 dated January 1986 using the methods ASTM D 2699 and D 2700, under international arrangements, in Standards TSO 5163 and 5164. The measurement method defined by these Standards is complex since it involves a special internal combustion engine, and requires meticulous control of various parameters for comparing the operation of the engine with the motor fuel under test and with a reference fuel. By definition, the method is precise since it is one that defines the values obtained, but it is not very repeatable, having a repeatability limit of 0.5 to 1.2 depending on the octane level determined. However, it is tedious and time-consuming to perform because of the equipment used and the number of operations that are necessary.

In research, and particularly in production operations, it would be desirable to determine octane numbers with sufficient precision, in a simpler, less tedious, and above all quicker and more repeatable manner with a repeatability varying from 0.1 to 0.2.

Correlations have been established between the chemical composition of a product in terms of its several components, and its octane number, the composition being determined by various means such as chromatography or nuclear magnetic resonance.

IR spectroscopic analysis is not generally used for determining the composition of petroleum products. If it is, the far infrared region is used, because the near infrared band is extremely complex for determining product compositions.

It is an object of the present invention to provide a simple, rapid and reliable method for the direct determination of the octane number without determining the chemical composition of the product undergoing examination.

According to the present invention there is provided a method for the direct determination of the octane number of a product from its near infrared (NIR) absorption spectrum in the wave number spectral range from 6667 to 3840 cm.sup.-1, preferably 4800 to 4000 cm.sup.-1 which method comprises selecting a number (n) of frequencies within this range and correlating the (n) absorbance values with octane number, the correlation being achieved by means of multivariate regression analysis.

The correlation depends on the type of spectrometer used, the type of octane number required, and the number (n) of the frequencies used.

(n) is suitably between 1 and 200 and preferably between 1 and 30.

A feature of the invention is the use of radiation in the near infrared, in the 1.5.mu. to 2.6.mu. band (6687 to 3840 cm.sup.-1), preferably the 2.1.mu. to 2.5.mu. band (4800 to 4000 cm.sup.-1), which, as explained above, is not normally used since it is made up of absorption bands that are combined in a complex manner; on the other hand, it ensures better repeatability for operating the method according to the invention.

Good results are obtained by using the following 15 frequencies expressed as the wave number (per centimeter), or frequencies of values close to these:

F.sub.1 =4670 cm.sup.-1 F.sub.2 =4640 F.sub.3 =4615 F.sub.4 =4585 cm.sup.-1 F.sub.5 =4485 F.sub.6 =4385 F.sub.7 =4332 F.sub.8 =4305 F.sub.9 =4260 F.sub.10 =4210 F.sub.11 =4170 F.sub.12 =4135 F.sub.13 =4100 F.sub.14 =4060 F.sub.15 =4040

In this case the baseline is taken at 4780 cm.sup.-1. The corresponding frequency expressed in practical units (Hz) is obtained by multiplying these values by 3.10.sup.10 --the velocity of light in cm/s.

An infrared spectrometer is used, e.g. Perkin Elmer model 1750, with a Fourier transform, provided with a 500 u sodium chloride cell in which a sample of the product is placed. Using the classic procedure, the absorbance, i.e. the logarithm of the ratio of the reduction between the incident radiation and the radiation after passing through the cell, is determined for each frequency.

The spectrometer used measures the absorbances ABS.sub.i, for the 15 significant frequencies F.sub.i selected, the octane number ON being calculated directly by means of a linear expression with a constant term C and appropriate multiplication coefficients A.sub.i :

ON=C+.SIGMA.ABS.sub.i .times.A.sub.i

Table 1 hereafter gives the values of these terms C and A.sub.i respectively, for the six following types of octane number:

clear research octane number (RON CLR) research octane number with 0.15 g/1 lead tetraethyl (RON 0.15) research octane number with 0.4 g/1 lead tetraethyl (RON 0.4) clear motor octane number (MON CLR) motor octane number with 0.15 g/1 of lead tetraethyl (MON 0.15) motor octane number with 0.4 g/1 lead tetraethyl (MON 0.4)

Other octane numbers can also be determined in the same way, for example, those of leaded gasolines with other lead alkyls, e.g. tetramethyl lead.

TABLE 1__________________________________________________________________________ RON CLR RON 0.15 RON 0.4 MON CLR MON 0.15 MON 0.4__________________________________________________________________________C 94.94 135.88 93.30 82.56 89.81 89.42A.sub.1 271.30 0 39.84 -204.35 -121.17 -231.42A.sub.2 -0.54 0 -69.56 382.25 0 453.95A.sub.3 -209.08 0 -105.05 -73.26 -107.81 -147.28A.sub.4 -14.24 0 75.88 -97.77 112.83 -18.16A.sub.5 16.51 46.52 20.93 2.39 -36.83 -107.71A.sub.6 28.84 47.80 37.80 21.41 76.71 73.12A.sub.7 26.05 0 30.14 -0.12 0 -13.56A.sub.8 16.28 -12.46 1.17 32.46 -1.01 54.12A.sub.9 16.03 -44.96 11.85 -43.20 -2.92 -66.38A.sub.10 -96.80 -106.30 -223.22 -118.47 -92.33 -250.18A.sub.11 -25.69 -11.59 41.77 0.65 -49.24 81.21A.sub.12 91.10 73.51 159.45 124.37 0 78.54A.sub.13 -141.96 0 -179.25 -71.57 -28.35 20.64A.sub.14 -27.62 -22.47 5.28 18.64 0 -9.39A.sub.15 56.30 0 74.70 21.50 76.57 33.64Motor/IR 0.212 0.151 0.193 0.252 0.291 0.273residualstandarddeviationIR 0.21 0.12 0.15 0.17 0.13 0.22Repeatability__________________________________________________________________________

If the spectrometer is equipped with means for calculation and a memory as in the type indicated, it is sufficient to program it and load the memory using the values given in the table. Otherwise, it can be connected to a computer which will carry out this operation.

By way of example, the results of the absorbance measurements obtained from 5 different gasolines are given in the upper part of the following Table 2. The values of the six corresponding octane numbers, as given by the method according to the present invention, are shown in the bottom part of the Table and compared with those in the mid part which were determined by the standard procedure.

TABLE 2__________________________________________________________________________ Gasoline 1 Gasoline 2 Gasoline 3 Gasoline 4 Gasoline 5__________________________________________________________________________F.sub.1 = 4670 0.06008 0.15214 0.09552 0.06692 0.10917F.sub.2 = 4640 0.06201 0.13162 0.09383 0.00681 0.10700F.sub.3 = 4615 0.08608 0.15967 0.13218 0.09455 0.15164F.sub.4 = 4585 0.06532 0.13226 0.09167 0.06753 0.10789F.sub.5 = 4485 0.07865 0.07443 0.05108 0.08144 0.06739F.sub.6 = 4385 0.52602 0.43641 0.49470 0.52258 0.49834F.sub.7 = 4332 0.84231 0.60741 0.73268 0.80791 0.73177F.sub.8 = 4305 0.68969 0.52745 0.62381 0.68184 0.64397F.sub.9 = 4260 0.63882 0.51533 0.53247 0.62656 0.56382F.sub.10 = 4210 0.41042 0.32350 0.36748 0.39421 0.37351F.sub.11 = 4170 0.41526 0.32331 0.36939 0.39696 0.36120F.sub.12 = 4135 0.38551 0.30219 0.33786 0.37445 0.33837F.sub.13 = 4100 0.39272 0.34255 0.35733 0.38637 0.36765F.sub.14 = 4060 0.54368 0.77413 0.60849 0.55774 0.64733F.sub.15 = 4040 0.44967 0.54314 0.53872 0.46912 0.56592MeasuredRON CLR 91.7 95.8 93.2 93.0 92.0RON 0.15 95.0 97.9 97.5 96.5 95.8RON 0.4 96.0 99.3 99.1 98.1 97.9MON CLR 78.2 82.2 83.9 79.6 81.8MON 0.15 81.0 84.2 87.7 82.3 85.9MON 0.4 82.5 85.2 92.3 84.1 88.6CalculatedRON CLR 91.4 95.8 93.3 92.8 92.5RON 0.15 95.0 97.2 98.0 96.5 95.9RON 0.4 95.9 99.3 99.2 97.6 97.7MON CLR 78.3 82.2 84.0 79.8 82.4MON 0.15 80.5 83.9 87.2 82.7 86.5MON 0.4 82.7 85.2 92.3 84.4 88.1__________________________________________________________________________

By comparing the results obtained by this method with those obtained by the conventional ASTM method, it can be shown that the precision obtained is very satisfactory.

The same method could, of course, be employed using slightly different frequencies F.sub.i or a number of them (n) greater or less than 15 (between 1 and 200), or again, with a different spectrometer, determining each time the different terms C and A.sub.i by multi-variable regression.

If the product is intended to be "leaded", the method gives, beforehand, the octane number which will be obtained after the addition of lead alkyls. The method can be used for gasolines whether or not they contain lead alkyls, the latter being used in amounts which are too low to affect the value of the absorbances. Thus, the correlation takes into account the susceptability of the gasoline to lead and the sensitivity, the sensitivity being defined for a given gasoline as the difference between RON and MON as determined by the previously described tests. This is valid whatever the amount of added lead. The method thus enables the clear research, clear motor, and leaded fuel octane numbers to be determined simultaneously.

The method is applicable to all types of internal combustion engine fuels whatever their composition as regards saturated, unsaturated and aromatic hydrocarbons. Suitable basestocks include straight run gasolines, steam-cracked gasolines, thermally-cracked or catalytically cracked gasolines, reformates, alkylates, hydrogenated gasolines and gasolines resulting from polymerisation, isomerisation, cyclisation, and dehydrogenation reactions, and other synthetic gasolines such as synfuels.

Where the product analysed contains oxygenated compounds used as anti-knock compounds, such as alcohols (e.g. t-butyl alcohol, methanol), aldehydes, ketones or ethers (e.g. methyl t-butyl ether), the method takes into account the fact that these compounds have been considered in establishing the statistical correlations.

Thus, the method according to the invention enables the determination, by means of a spectrometer, optionally equipped with a fibre optics system and optionally linked to a computer, of the octane number(s) which are of interest, virtually in real-time, and continuously.

In production, one application is the optimisation of the control of a production unit by measuring the octane number of the product and adjusting the feedback control of the unit in order to obtain a product with the desired octane number. This application requires the use of a process computer which may or may not be involved in the control system, for example, at the outlets of the processing units producing the gasoline basestocks (reformers, thermal, catalytic and steam crackers, etc.).

Another application lies in automatically rendering the mixing process non-interactive by determining the octane number of each storage tank in order to determine, by calculation, the proportions of the products issued from the various tanks for transferring to the mixing tank and/or determining the octane number of the product obtained.

Claims

1. In a process for optimizing the control of a production unit for making an internal combustion engine fuel product or base stock with a desired octane number, the improvement which comprises: (a) determining the octane number of said fuel product or base stock by (i) measuring the near infrared absorption of said fuel product or base stock within the wave number spectral range 4800-4000 cm.sup.-1, and (ii) directly correlating by means of multivariate regression analysis the (n) absorbance values obtained with the octane number of said product or base stock; and (b) employing said determined octane number to control and optimize the process to produce said desired octane number.

2. The process according to claim 1 wherein (n) is in the range 1-200.

3. The process according to claim 2 wherein (n) is in the range 1-30.

4. The process according to claim 3 wherein the (n) frequencies used are selected from those defined by the following:

F.sub.1 =4670 cm.sup.-1

F.sub.2 =4640 cm.sup.-1

F.sub.3 =4615 cm.sup.-1

F.sub.4 =4585 cm.sup.-1

F.sub.5 =4485 cm.sup.-1

F.sub.6 =4385 cm.sup.-1

F.sub.7 =4332 cm.sup.-1

F.sub.8 =4305 cm.sup.-1

F.sub.9 =4260 cm.sup.-1

F.sub.10 =4210 cm.sup.-1

F.sub.11 =4170 cm.sup.-1

F.sub.13 =4100 cm.sup.-1

F.sub.14 =4060 cm.sup.-1

F.sub.15 =4040 cm.sup.-1

5. The process according to claim 1 wherein the correlations are not affected by a susceptibility of the product to lead alkyls.

6. The process according to claim 1 wherein the product is a gasoline.

7. The process according to claim 7 wherein the gasoline contains lead alkyls.

8. The process according to claim 6 whereof the gasoline contains oxygenated compounds.

9. The process according to claim 1 wherein at least one of the following octane numbers are determined simultaneously: clear research, clear motor, leaded research and leaded motor.

10. A process as claimed in claim 1, wherein the production unit is selected from the group consisting of reforming, thermal, and catalytic cracking units.