JPH0692545B2 - Carbon black process control method and apparatus - Google Patents

Description

FIELD OF THE INVENTION The present invention relates to process control, and more particularly to process control for controlling carbon black production.

[Problems to be Solved by Prior Art and Invention]

In the production of carbon black, it is desirable to control certain output variables of carbon black to produce a substantially constant amount of carbon black. Carbon black output variables that are often the focus of control are iodine value and DBP. Since the input variables and other physical parameters of the process of carbon black production often change during the production of carbon black,
Producing a substantially constant amount of carbon black has proven difficult. Input variables that often fluctuate during the carbon black production process include, for example:
Air humidity and fuel quality. Fluctuations in input variables can significantly affect carbon black input variables such as iodine number and / or DBP. Similarly, other non-measurable physical parameters change frequently during the carbon black production process and also affect the output variables of the carbon black, such as iodine number and / or DBP.

In some known carbon black production systems, samples of carbon black produced are taken at spaced intervals, eg, once every few hours of operation.
Then output variables, eg iodine value and / or DB
P is measured for each sample. Then, after testing each sample, one or more input variables, such as the feedstock flow rate, are adjusted. This adjustment causes the output variables, such as iodine number and / or DBP, to return to their target values, usually based on subjective experiments using a particular carbon black production system.

One problem with using such known methods of controlling the production of carbon black is that the output variables of carbon black, iodine number and / or DBP are not controlled in the time interval between samples. Therefore, the input variables or other physical parameters of the carbon black production system are related to the output variables such as iodine value and / or
Alternatively, move the DBP value out of the desired range and this change will usually not be seen until the next sample is taken. As a result, substantial amounts of carbon black produced may not fall within customer specifications. Yet another problem in using such known methods of controlling the production of carbon black is to adjust one or more output variables based on the values of the output variables measured in the laboratory. As such, the method is like relying on the sovereign analysis of the operator. After all, the adjustment of the input variables can often change between operators and thus the amount of carbon black produced is not constant.

Therefore, it is an object of the present invention to overcome the problems and drawbacks of known carbon black production systems.

[Means for Solving the Problems, Actions and Effects of the Invention]

The present invention relates to a method of controlling the production of carbon black in a carbon black reactor, the method comprising the steps of: (a) at intervals of carbon black production during operation of the carbon black reactor. Measuring at least one input variable utilized, and (b) using at least one algorithm, the output of at least one carbon black based on the at least one input variable measured during the spacing interval. Predicting variables at spaced prediction intervals, (c) determining an average value of the at least two predicted outputs at spaced average intervals, and (d) at spaced intervals Between the mean value of at least one predicted output variable and the target value of the at least one output variable while the reactor is operating. By utilizing the difference, is achieved by adjusting at least one of the input variables, the target value of the output variable in order to obtain a carbon black substantially constant amount in accordance with regulation algorithm consists.

The method of the present invention preferably further comprises step (a) sampling the carbon black produced during the operation of the carbon black reactor at sampling intervals spaced apart by (b) the carbon black reactor During operation, measuring the at least one output variable predicted by the algorithm from the carbon black sample, and (c) the at least one algorithm based on the measured value of the at least one output variable. To better predict the at least one output variable.

In one embodiment of the invention, the at least one predicted output variable is iodine number and the at least one adjusted input variable is feedstock flow rate. In another embodiment of the invention, the at least one predicted output variable is DBP and the adjusted input variable is the potassium addition solution flow rate.

Another embodiment of the invention, at least one algorithm, is a weighted average of the variances of the errors in the predicted values of the carbon black output variables during the sampling period of the carbon black, and the measured values of the output variables. It is adjusted by taking advantage of the variance of the error. The best estimate of the output variable is based on ethylmethacrylate of the variance of the error and the difference between the measured value of the output variable and the predicted mean value of the output variable during the sampling period.

The invention also relates to an apparatus for controlling the production of carbon black in a carbon black reactor. The apparatus comprises metering means for measuring at least one input variable utilized in the production of carbon black at spaced intervals during operation of the carbon black reactor. The calculation means of the device are coupled to the metering means and determine at least one output variable of the carbon black according to at least one algorithm using the at least one input variable measured during the interval. Predict at intervals. The calculating means further determines the average value of the at least one predicted output variable at the spaced average intervals. The device further comprises adjusting means, which is coupled to the calculating means and adjusts at least one input variable of the carbon black at spaced intervals according to an adjusting algorithm. This adjustment achieves the target value while the reactor is operating based on the difference between the average value of the at least one predicted output variable and the target value of the output variable over the spaced average interval. To obtain a substantially constant amount of carbon black.

In another embodiment of the present invention, the apparatus further samples the carbon black produced during operation of the carbon black reactor at spaced intervals, thus setting said at least one output variable to carbon black. Sampling means that enables measurement from the sample. The computing means is responsive to the measured value of the at least one output variable to adjust the at least one algorithm by utilizing said measured value of the at least one output variable to more accurately predict the output variable.

Accordingly, the method and apparatus of the present invention compensate for changes in the input variables and other physical parameters of the carbon black production system while the carbon black reactor is operating to provide a substantially constant amount of carbon. Produces black. Outputs that measure at least one input variable at spaced intervals, predict at least one output variable with an algorithm that uses at least one input variable at spaced intervals, and predict at mean spaced intervals By averaging the variables and then adjusting the at least one input variable at spaced intervals utilizing the predicted value of the average of the output variables, the method and apparatus of the present invention provides a substantially constant amount of carbon black. To generate. Similarly, by sampling carbon black produced at spaced intervals, measuring at least one output variable from the sampled carbon black, and using the measured value to adjust at least one algorithm. The apparatus and method of the present invention accurately predicts output variables and thus produces a substantially more consistent amount of carbon black.

Other advantages and features of the invention will be apparent in view of the following detailed description and the drawings in combination therewith.

〔Example〕

The carbon black process control system of the present invention is
By compensating for variables in the physical parameters in the carbon black reactor by adjusting one or more input variables of this process, one of this process
Alternatively, two or more output variables are controlled, thus producing a substantially constant amount of carbon black. The carbon black output variables that are controlled are, for example, iodine number and / or DBP.

In FIG. 1, an example of a furnace carbon black reactor in which the process control system of the present invention can be used is shown schematically. The carbon black reactor shown is a three stage reactor including a burner zone, a feedstock injection zone, and a reactor zone.
However, the process control system of the present invention is used with other types of carbon black reactors or processes that pyrolyze hydrocarbon feedstocks with hot combustion gases to produce combustion products containing particulate carbon black. be able to. In the burner zone of the reactor in FIG. 1, liquid or gaseous fuel is reacted with any type of oxidant, preferably air, to form a hot combustion gas. The resulting combustion gases exit the downstream end of the burner zone and flow at high velocity through the feedstock injection zone. The hydrocarbon feedstock in the form of a gas, vapor or liquid can be the same as or different from the fuel utilized to form the combustion gas stream and is injected into the combustion gas stream in the feedstock injection zone. However, thermal decomposition of the hydrocarbon feedstock occurs. The reaction mixture of feedstock and combustion gases is then discharged into the reaction zone,
Here, the formation of carbon black particles is completed. The reaction mixture is then quenched at the end of the reaction zone with a suitable fluid, usually water, to quench the carbon black particle formation reaction. The reaction mixture is then cooled further,
The solid carbon black particles are then collected by methods known in the art.

The input variables analyzed by the carbon black process control system of the present invention are also shown schematically in FIG. Each of the input variables is measured prior to injection into the burner zone or feedstock injection zone.
Input variables include: feedstock flow rate, fuel flow rate, air flow rate, air preheat temperature, air humidity, gas or other first stage fuel quality, feedstock quality, And / or the flow rate of the potassium addition solution. Usually, only some of the input variables can be precisely controlled to control one or more output variables of the carbon black, such as iodine number and / or DBP. Typical controlled variables are feedstock flow rate, fuel flow rate, air flow rate, and / or potassium addition solution flow rate.

In one embodiment of the invention, the process control system calculates the predicted iodine number (I ₂ No. _P ) at spaced intervals, eg, every 1-10 seconds. The predicted iodine number is calculated by an algorithm that is based in part on experimental results for a given carbon black reactor geometry using a process control system.
Based on. The predicted iodine values are then averaged over intervals, eg, every 2 minutes (I ₂ No.
_AVG ). Based on the average predicted iodine number, a controlled input variable, eg, feedstock flow rate, is automatically adjusted to achieve the target iodine number (I ₂ No. _GOAL ). Therefore, a substantially constant amount of carbon black is a measurable input variable of the carbon black reactor,
For example, it can be generated independently of changes in air humidity and / or calculated input variables, such as changes in fuel quality.

In accordance with an embodiment of the present invention, the process control system is used with a three stage reaction as shown schematically in FIG. An exemplary reaction uses a hydrocarbonaceous oil feedstock and natural gas fuel. However, the process control system of the present invention may be used equally well with other types of reactor geometries, and other types of feedstocks and / or fuels. The predicted iodine value (I ₂ No. _P ) can be calculated according to the following iodine value algorithm: (1) I ₂ No. _P = KC * OAC + KP * PC + KA * AIR + KT * CAT + K
The constants for the H * AH + KO algorithm are determined empirically for a given carbon black reactor geometry. For example, the first
As shown, the algorithm for the three-stage reactor can have different values than the constants of the algorithm for the two-stage reactor (not shown). The constants of the algorithm are defined as follows: KC-constant of overall combustion KP-constant of primary combustion KA-constant of air velocity KT-constant of air preheating temperature KH-constant of air humidity KO- System Concept Constants The forward feed input variables are defined as: OAC-Overall Combustion [%] PC-Primary Combustion [%] AIR-Combustion Air Velocity [KSCFH] CAT-Combustion Air Preheat Temperature [° F] AH-Absolute Air Humidity [pounds of water / 1000 pounds of dry air] The input variables for the forward feed are certain input variables of the carbon black reactor while the reactor is operating. , Determined by measuring with a weighing device. As soon as each input variable is measured, the forward-feed input variable is calculated based on the following equation: Where AIR is the air flow rate [KSCFH] (standard cubic feet / hour, 1000), GAS is the gas flow rate [KSCFH], and ATBG is the air to combustion gas ratio [SCF air / SCF gas]. Which is the stoichiometric value of the amount of air required to completely combust the corresponding volume of gas.

If the carbon black reactor uses a fuel of a type other than gas, such as a liquid hydrocarbon fuel, the fuel flow rate is given in equation (2) instead of gas flow rate (GAS), and this time also appears below. It will be shown in other equations. Similarly, ATBG would be replaced in the same equation as the ratio of the stoichiometric value of the amount of air required to completely burn out the corresponding amount of the type of fuel used. Similarly, if the carbon black reactor uses a suitable oxidant other than air, the flow rate of the oxidant is shown in equation (2), and in other equations described later that time also appears (AI).
R) will be shown instead.

Where: AIR is the air flow rate [KSCFH] (standard cubic feet / hour, 1000), GAS is the gas flow rate [KSCFH], ATBG is the air to combustion gas ratio [SCF air / SCF gas] , OIL is the liquid hydrocarbon feedstock flow rate [gallons / hour], and ATBO is the air-to-combustion oil ratio [KSCF air / gallon oil], which gives the corresponding volume of oil completely. It is a stoichiometric value of the amount of air required to burn (typical value is about 1.54 KSCF / gallon oil).

If the carbon black reactor uses a feedstock other than a liquid hydrocarbonaceous feedstock, such as a gaseous hydrocarbon feedstock, the feedstock flow rate may be changed at this time instead of the oil feedstock flow rate (OIL). Will also appear in other equations described below. Similarly, ATBO will be replaced in the same equation as the ratio of the stoichiometric value of the amount of air required to completely combust the corresponding amount of feedstock of the type used.

Air flow rate (AIR) and gas flow rate (GAS) are measured online by known instruments prior to injection into the burner zone of the carbon black reactor. The air and gas meters are preferably orifice-type meters that compensate for pressure and temperature variables that flow when generating the flow rate signal. ATBG is preferably calculated based on the input gas composition measured by a gas chromatograph (not shown). Gas chromatographs can be used to determine gas composition periodically on-line or periodically off-line. Based on the current gas composition, the ATBG value is adjusted accordingly. Similarly, the specific gravity of the gas used by the gas meter is also adjusted accordingly based on the gas composition readings of the gas chromatograph. When a gas chromatograph measures gas composition online, it usually has the ability to update ATBG values within a range of at least about 2-10 minutes. On the other hand
ATBO is usually not measured and cannot be updated online. Therefore, ATBO values are preferably measured in the laboratory for each particular grade of feedstock or blend of feedstocks. The ATBO value can be updated, for example, before the production run or once every few months.

The flow rate (OIL) of the feedstock is preferably measured by a Coriolis type flow meter,
This flow meter measures feedstock mass flow rate, usually in pounds per hour, and feedstock density.
Measure before injection into the injection zone of the reactor. The feedstock flow rate is preferably converted to a corrected volume flow rate and is expressed in gallons / hour. The preheat temperature (CAT) of the combustion air is measured by a thermocouple just before entering the reactor barber zone. Absolute humidity of air (A
H) is measured by a humidity sensor of the type known in the art and is expressed in units of pounds of water / 100 pounds of dry air. The absolute humidity measurement of air is preferably used for two purposes. One purpose is to provide an updated forward feed input variable (AH) for the iodine number algorithm. Another purpose is to depend on the measured absolute humidity of the air (AH), the air velocity (AI
R) is adjusted to maintain a substantially constant dry air flow rate into the burner zone of the reactor. Preferably, a PID algorithm of the type known in the art (Proportional, Integral, Inductive control algorithm) is used to adjust the air flow rate depending on the updated absolute humidity reading of the air, It compensates for the amount of humidity in the air, thus maintaining a substantially constant dry air flow rate.

The algorithmic constants of the iodine number algorithm (Equation (1)) are determined according to known process identification procedures using regression analysis, and for a given type of carbon black reactor geometry. Therefore,
The values of the constants appear to differ for substantially different reactor geometries. Known software packages (includes the following components: "RS / 1", "RS / Explorer"
e ”and“ RS / Discover ”, BBN Sftware Products Co
Rporation, sold by Cambridge, Mass.) is preferably used to perform the regression analysis procedure. BBN software can be used with a VAX microcomputer manufactured by Digital Equipment Corporation, Minard, MA. BB
N software facilitates and is not necessarily required to perform experimental design procedures known in the art, and also regression analysis procedures known in the art, for performing such procedures. It simply provides a convenient means.

When performing the regression procedure, the input and output variables are identified in the carbon black production process. The input variables relating to the iodine value are, for example, those shown in FIG. 1, and the feed stock flow velocity, the air flow velocity,
Includes fuel flow rate, air preheat temperature and humidity, fuel quality (ATBG), and feedstock quality (ATBO). The output variable is the iodine value (I ₂ No.). Based on the input and output variables identified, a series of experiments is designed to identify the algorithm parameters, preferably by using BBN software on a VAX minicomputer. This series of experiments is then carried out in a carbon black reactor with reactor geometry of the type for which the algorithm is used. Therefore, the regression analysis means will provide constants having different values for the geometry of different types of reactors.
At different stages during the experiment, the input variables are changed in the manner defined by the designed experiment. Based on experimentation, collect a set of input and corresponding output data.
A regression analysis procedure is then performed on the data set to identify the experimentally determined constants of the iodine number algorithm (Equation (1)).

In one embodiment of the present invention, following the regression analysis procedure described above, the following constants were experimentally determined for a three stage reactor geometry similar to that shown schematically in FIG. : KC = 12.5 KT = 0.094 KP = −0.123 KH = −0.238 KA = −0.184 KO = −201 (approximate) Therefore, according to one embodiment of the present invention, a forward supply of the iodine value algorithm (Equation (1)). The input variables necessary for the determination of the input variables are measured about once per second. Then, based on those measurements, the algorithm for iodine value was solved about once per second to obtain a new predicted iodine value (I ₂
No. _P ) is generated. The predicted iodine values calculated over the intervals are then averaged over the intervals (I ₂ No. _AVG ), eg, about every 2 minutes. The controlled input variable, eg, feedstock flow rate (OIL), is then adjusted to the average predicted iodine value (I ₂ N
o. _AVG ) up to the set point of iodine value (I ₂ N
o. _GOAL ) is adjusted automatically at the end of each mean interval depending on the difference between the two to achieve the target iodine value. However, one or more other input variables, such as AIR and / or GAS, may be adjusted instead of the feedstock flow rate (OIL) to achieve the target iodine number (I ₂ No.
_GOAL ) can be achieved.

The relationship between iodine value and OAC is a first-order regulatory relationship. O
AC is the calculated control variable as opposed to the measured control variable. As described below, the equations defining OAC include AIR, GAS, and OIL as they are now. Therefore,
Based on the relationship between iodine value and OAC, the desired measured control variable, the target iodine value (I ₂ No. _GOAL ), which induced an appropriate change in OIL, can be achieved. Feedstock flow rate (OIL) is the preferred input variable to control. Because, for one reason, it only appeared around the time of the iodine number algorithm, so the regulation procedure is relatively simple and straightforward.

The new feedstock flow rate (OIL _NEW ) required to achieve the target iodine value I ₂ No. _GOAL is
Estimated based on the relationship with OAC: (4) ΔI ₂ No. = KC * ΔOAC where: ΔI ₂ No. is the average of two minutes of I ₂ No. _GOAL −I ₂ No. _P (or Other intervals) (I ₂ No. _AVG ), ΔOAC is I ₂ No. _GOAL- average of 2 minutes of measured OAC (OAC
A new OAC (OAC _NEW ) required to achieve _AVG ),
And KC is the overall burning constant of the iodine number algorithm.

Adapted from the partial derivation of the iodine value algorithm for OAC (Equation (1)), such as Equation (4). The flow rate (OIL _NEW ) of the new feedstock is then determined based on the following equation: Then solve equations (5) and (6) for OIL _NEW as follows: Therefore, OIL _NEW calculates every two minutes (or at other intervals) using the predicted average iodine value (I ₂ No. _AVG ) calculated over that average interval and then the feedstock. Flow rate (OIL) of I ₂ N
o. _GOAL can be achieved.

The carbon black process control system of the present invention is
An additional feature is to have an offline laboratory measurement procedure. At intervals, while the carbon black reactor is operating, samples of the carbon black produced are taken and the iodine number of each sample is determined by known techniques (I ₂ No. _LAB ). The measured iodine value (I ₂ No. _LAB ) and its known standard deviation (SD _LAB ) are the mean and standard deviation (I ₂ No. _P ) of the expected iodine value (I ₂ No. _P ) during the sampling period. Determine with SD _P ). Then, the measured iodine value (I ₂ No. _LAB ) value, the standard deviation of the test (S
D _LAB ), and the expected iodine value (I ₂ No. _P ) mean and standard deviation (SD _P ) depending on the system's intercept constant (KO) of the iodine value algorithm (Equation (1)). Adjust to calculate a more accurate predicted iodine number (I ₂ No. _P ), as described in more detail below.
Thus, according to the present invention, the accuracy of the iodine number control algorithm (Equation (1)) is systematically checked against the iodine number (I ₂ No. _LAB ) measured in the laboratory, and the carbon black reactor is Improved while running. Thus, the off-line sampling feature of the present invention either compensates for unmeasured perturbations to the carbon black reactor that are not accurately measured, or, as previously mentioned, can be measured in contrast to measurable input variables. Can not.

In accordance with the present invention, a filter algorithm, preferably the Kalman filter algorithm, is used to intercept the iodine value algorithm system intercept (K
O) is applied to change. The system's intercept (KO) varies based on the measured iodine value (I ₂ No. _LAB ) and the expected iodine value (I ₂ No. _P ) determined during the carbon black sampling interval. Let's make an algorithm of iodine value that predicts iodine more accurately. Iodine number (I ₂ N
_o.LAB ) is measured by methods known in the art, for example, by the volumetric method in which a sample of carbon black is titrated with an iodine solution. The iodine value test is preferably carried out according to the iodine adsorption number given by the ASTM designation: D1510-85. The sampling interval during operation of the carbon black reactor is usually in the range of about 2-20 minutes.

According to the sampling features of the present invention, the best estimate of the current expected variance of the iodine value error (V _IP ) and the laboratory measured variance of the iodine value error (V _IL ) are determined. The variance of the error is the square of the standard deviation of the iodine value. Therefore, V _IL is the square of the standard deviation of iodine number (SD _LAB ) measured in the laboratory for a sample of carbon black. Usually only one laboratory measured iodine value (I ₂
No. _LAB ) is taken during each sample period, so V _IL is essentially a constant that is not accurate for iodine numbers measured in other laboratories of the type known in the art. Or determined by reproducibility studies. Therefore, V
_ILs are usually updated, eg, once every few months, or when changes are dependent on the procedure to determine the iodine value (I ₂ No. _LAB ) measured in the laboratory. V _IL is the best estimate of the current predicted iodine number (I ₂ No. _P ), as described in further detail below. Thus, V _IP and V _IL are each an indication of uncertainty in their respective determined iodine value.

Based on the variances of the errors V _IP and vvil, the Kalman filter increase (K _I ) of iodine value is then used to update the system intercept (KO) of the iodine value algorithm to To decide: Therefore, the Kalman filter increase (K _I ) is essentially a weighted average of the error variances (V _IP and V _IL ),
This is a measurement with two normal noises each (I ₂ No. _P and
I ₂ No. _LAB ) reflects the degree of fluctuation. I ₂ No. _P and I ₂ No. _LAB are usually different. Therefore, the Kalman filter increase (K _I ) is actually two different measurements, I ₂ No. _P and I ₂ No., which indicate which measurement is more accurate.
It is a weighting coefficient based on statistical information about _LAB reliability. For example, if K ₁ = 1 then there is a negligible error variance in I ₂ No. _LAB , and if K _I = 0 then there is a negligible error variance in I ₂ No. _P. To do.

Based on the Kalman filter increase (K _I ) and using the Kalman filter algorithm, the new optimal estimated iodine number (I ₂ No. _FILTER ) is
(9) I ₂ No. _FILTER = I ₂ No. _AVG + K _I * (I ₂ No. _{LAB −} I ₂ No.
_AVG ) where I ₂ No. _AVG is the average of the predicted iodine value (I ₂ No. _P ) during the sampling period.

Then a new optimal estimated iodine value (I ₂ No.
Based on _FILTER ), the intercept constant (KO _NEW ) of the new system for the iodine number algorithm is calculated as follows: (10) KO _NEW = KO _OLD + I ₂ No. _FILTER −I ₂ No. _AVG 1 A change in point corresponds, for example, to a change in iodine value at one point in the intercept constant (KO) of the system, so the number can be directly substituted into equation (10) to obtain KO
You can solve about _NEW . Therefore, the intercept constant (KO) of the system can be adjusted to make the iodine value algorithm (Equation (1)) more accurate using the iodine value (I ₂ No. _LAB ) measured each time in the laboratory. become able to.

Referring again to the error variance, the best estimate of the true current error variance (V _IP (k + 1)) of the predicted iodine number in the time interval (k + 1) is, as described below, the Kalman filter It is used to determine the increase (K _I ) and is determined as follows: (11) V _IP (k + 1) = V _IE (k) + V _IM (k + 1) where: V _IP (k + 1) is the time Is the best estimate of the current predicted iodine value (I ₂ No. _P ) in the interval (k + 1), and V _IE (k) is the previous optimal iodine value estimate in the time interval (k). (I ₂ No. _FILTER is the variance of the error, and V _IM (k + 1) is the predicted iodine value (I ₂ ) in the time interval (k + 1) measured over the last sample period.
No. _P ) error variance.

Then increase the new Kalman filter (K ₁ (k +
1)) is determined from the variance of the error between the current predicted iodine value (I ₂ No. _P ) and the current laboratory measured iodine value (I ₂ No. _LAB ) as follows: : V _IL (k + 1) is the iodine value (I
₂ No. _LAB ) error variance, and is defined as: (13) V _IL (k + 1) = [PSD _LAB / 100] ² * I ₂ NO. _GOAL PSD _LAB % Test standard deviation of iodine number, determined by precision or reproducibility studies known in. Therefore, the new optimal Kalman filter increase (K _I (k + 1)) is substituted into equation (9) above to solve for the new optimal predicted iodine number (I ₂ No. _LAB ). Then substitute I ₂ No. _LAB into the above equation (10) to obtain the intercept constant of the new system (KO _NEW ).
So that the iodine value algorithm predicts more accurately than the iodine value.

Then at the end of the next sample period (the new optimal estimated iodine value error variance (V _IE (V _IE (k) V _IP (k + 1) in equation (11) above should be used in the determination of V _IE (k) V _IP (k + 1) k + 1)) is determined as follows: In accordance with another embodiment of the present invention, a process control system is used to control the structure of carbon black. The structure of carbon black is normally laboratory, ASTM designation: D2414-86
The absorption value of dibutyl phthalate given by (“DB
P ”). Therefore, the DBP value is an indication of the structure of carbon black. However, there are other suitable measures of carbon black structure that can be equally controlled by the process control system of the present invention. One method of DBP control is to inject a potassium addition solution (K ⁺ S) known in the art, preferably into the feedstock, prior to injecting the feedstock into the reactor's feedstock injection zone. By doing. The potassium addition solution (K ⁺ S) is then dispersed in the reaction mixture in the reaction zone, which has the effect of ionic charge on the particles of carbon black thus formed. Therefore, when a higher concentration of potassium loading solution (K ⁺ S) is injected into the feedstock, there will usually be less tendency to agglomerate between the particles of carbon black formed.

In accordance with the present invention, the predicted DBP value (DBP _P ) is calculated at spaced intervals, eg, every 1-10 seconds. The predicted DBP value (DBP _P ) is calculated by the DBP algorithm, which is based in part on experimental test results for a given carbon black reactor geometry using a process control system. The predicted DBP values are then averaged over spaced intervals, eg, every 2 months (DBP _AVG ). Average predicted DBP value (DBP _AVG )
A controlled input variable, for example, the flow rate of the potassium addition solution (K ⁺ S) is automatically adjusted to achieve the target DBP value (DBP _GOAL ).

The predicted DBP value (DBP _P ) can be calculated according to the following DBP algorithm: (15) DBP _P = (164.9-17.3 * X) * F [for O <X <1] and (16) DBP _P = (147.6-17.3 * In (X)) * F [X> 1
About] where: X is the concentration of potassium ions (K ⁺ ) in the feedstock [g K ⁺ / 100 gallons of oil], and F is for unmeasured perturbations in the carbon black reactor, or The difference between the responses is a scale factor calculated to control the algorithm (F is usually in the range of about 0.7 to about 1.2).

The constants in the DBP algorithm use regression analysis according to the known process identification procedure for a given carbon black reactor geometry in the same manner as described above for the determination of the algorithm constants for the iodine number algorithm. Experimentally determined. Therefore, the value of the constant appears to be different for different types of reactor geometry. The measured input variables for DBP are preferably the potassium addition solution flow rate and the feedstock flow rate. The output variable is DBP or other suitable measure of carbon black structure. A series of experiments is then carried out in a carbon black reactor with reactor geometry of the type using the algorithm, as described above for the iodine number algorithm. Based on the experiment, 1
Collect data for a set of inputs and reacting outputs. The regression analysis procedure is then performed on that set of data and the DB
Identify the P algorithm constants. The constants in the DBP's algorithm defined in Eqs. (15) and (16) were experimentally determined for a three stage reactor similar to that shown schematically in FIG. 1 according to the regression analysis procedure described above.

DBP Algorithms, Equations (15) and Equations (16)
Using, for example, 1 at 1 second intervals
Predict the DBP value (DBP _P ). The predicted DBP values are then averaged over a spaced average interval, for example once every two months (DBP _AVG ). Then the average DBP value (DBP
_AVG ), using the DBP regulation algorithm defined as
Calculate ⁺ S _NEW ): (17) K ⁺ S _NEW [pounds / hour] = RATIO [pounds K ⁺ S / gallons of oil] * OIL _NEW [gallons / hour] where: X _NEW is the potassium ion concentration (X) in the feedstock
Derived from the partial derivation of the DBP algorithm (equations (15) and (16)) with respect to and defined as: K _MIX is the intensity of a mixture of potassium additive solution (K ⁺ S), which is g of potassium ions per pound of potassium additive solution ^{(K + S) (K +} ). X _NEW is potassium ion (K ⁺ ) in the feedstock required to achieve DBP _GOAL
Is a new concentration of.

K ⁺ S _AVG is the average potassium loading solution flow rate during the 2 month interval and OIL _AVG is the average feedstock flow rate during the 2 month interval. OIL _NEW is the current flow rate set point for the feedstock, which is preferably adjusted according to the iodine number algorithm as described above. Therefore, by utilizing the average predicted DBP value (DBP _AVG ) over a 2-minute interval, the flow rate (K ⁺ S _NEW ) of the new potassium addition solution was determined according to equation (17) and the target DBP value was determined. (DBP _GOAL ) can be achieved.

The process control system of the present invention has an additional feature, an offline DBP laboratory measurement procedure. Samples of carbon black produced are taken at intervals, while the carbon black reactor is operating, and the DBP for each sample is determined by methods known in the art.
Measure the value (DBP _LAB ). The sampling interval when collecting carbon black samples is usually within the range of about 2 to 20 minutes. The DBP _LAB is preferably an AS, as described above.
TM display: Measure according to D2414-86.

The measured DBP value (DBP _LAB ) and its known standard deviation (S
D _LAB ) is the estimated DBP value (D
Mean and standard deviation of BP _P) (determined with SD _P).
Next, the measured DBP value (DBP _LAB ) and its standard deviation (SD
_LAB ), and the mean and standard deviation of the predicted DBP value (DBP _P ), depending on the DBP algorithm (Equation (15)
And adjust the factor (F) on the scale of (16)) to calculate a more accurate DBP value. Thus, in accordance with the present invention, the accuracy of the DBP algorithm itself is determined by the DB measured in the laboratory while the carbon black reactor is operating.
It can be systematically tested and refined for P-value (DBP _LAB ).

In accordance with the sampling features of the present invention, the variance of the predicted DBP value error (V _DP ) and the laboratory measured error of the DBP value (V _DL ) are determined. V _DL measured in the laboratory
It is the square of the standard deviation of the DBP value (DBP _LAB ). Since preferably only one laboratory measured DBP value is taken during each sample period, V _DL is essentially a constant, which is a type of DBP _LAB measurement procedure known in the art. Another precision or reproducibility of. Therefore,
V _DL is typically updated periodically, eg, once every few months, or whenever there is a change in the DBP _LAB decision procedure. V _DP is the best estimate of the variance of the error of the current predicted DBP value (DBP _P ), as described in more detail below.

Determine the best estimate of the true DBP value (DBP _FILTER ) during the sampling period using a filter algorithm, preferably the Kalman filter algorithm, based on the variances of error V _DP and V _DL To do. DBP _FILTER is
It is generated as a weighted average between the DBP _LAB and the average of the predicted DBP values during the sampling period. (DBP
Kalman filter algorithm DBP for _FILTER) is defined as _{follows: (22) DBP FILTER = DBP} AVG + K D * (DBP LAB -DBP AVG) K D is the increase in the Kalman filter DBP, This is essentially a weighted average of the error variances V _DP and V _DL ,
And is defined as: Then, based on the DBP _FILTER , adjust the scale factor (F) of the DBP algorithm, equations (15) and (16) (F _NEW ), and the DBP algorithm makes the DBP more accurate. Let us predict: V _AVG is the average concentration of potassium addition solution (K ⁺ S) in the feedstock, as defined in equation (21), during the sampling period. Then a new scale factor (F _NEW ) is substituted into the DBP algorithm (Equations (15) and (16)) to replace the previous scale factor (F), thus adjusting the algorithm to be more accurate than DBP. Predict.

Best estimate of the variance of the true current error of the predicted DBP _{value in} the time interval (k + 1) (V _DP (k + 1))
Is used in equation (23) to determine the Kalman filter increase (K _D ) of the current DBP and is defined as: (26) V _DP (k + 1) = V _DE (k) + V _DM (K + 1) where: V _DP (k + 1) is the best estimate of the true current predicted DBP value in the time interval (k + 1), and V _DE is the previous optimal DBP value in the time interval (k). Is the variance of the error in the estimate of (DBP _FILTER ), and V _DM (k + 1) is the predicted DBP value (DB) in the time interval (k + 1) measured over the last identical period.
It is the variance of the error of P _P ).

Then increase the Kalman filter of the new DBP (K
_D (k + 1)) is the current predicted DBP value (DBP _P ) and the current laboratory measured DBP value (DB
P _LAB ) determined as the weighted average of the variances of the errors: V _DL (k + 1) is the DBP value (DB
The error variance of P _LAB ) and is defined as: (28) V _DL (k + 1) = [PSD _LAB / 100] ² * DBP _GOAL where PSD _LAB is known in the art. Is the existing standard deviation of the DBP in the laboratory, as determined by the accuracy or reproducibility study. Therefore, substituting the Kalman filter increase (K _D (k + 1)) of the new DBP into equation (22) above, we obtain the new optimal estimated DBP value (DBP
_FILTER ). The DBP _FILTER is then substituted into equation (24) or (25) above to cause DBP's algorithm (Equations (15) and (16) to predict DBP more accurately.

Then the period of the next sample (V in equation (26) above
New optimal estimated DBP value error variance (V _DE to be used when determining V _DP (k + 1) in _DE (k))
(K + 1)) is determined as follows: In another embodiment of the invention, the process control system further comprises a controlled output variable, for example CUSUM (“cumulative sum”), for monitoring iodine value and / or DBP. CUSUM is an iodine value algorithm, D
It is not completely compensated by the BP algorithm or the respective Kalman filter algorithms. It compensates for trends in iodine number or DBP, which can be the result of unmeasured confusion in the carbon black reactor. Therefore, CUSUM monitors the I ₂ No. _LAB and measures each output variable to determine if there is a shift in the mean of values sufficient to require further adjustment in this process. Every time, CUSUM monitors DBP _LAB .

Each CUSUM uses the two higher totals (S _{H (i)} ) and the lower total (S _{L (i)} ) of the cumulative totals, respectively, to calculate I ₂ No.
Test the _LAB and DBP _LAB to determine if there are any unwanted trends. When the running totals are reset, set each running total ( _{SH (i)} ) and ( _{SL (i)} ) equal to zero. Two sums is determined as _{follows: (30) S H (i} ) = Max [O, S H (i-1) + Y i - (GOAL + k)] (31) S L (i) = Min [ _{O, S L (i-1} ) + Y i - (GOAL-k)] where: S _{H (i-1),} since the last cumulative total is reset, the sum total of all previous high side And S _{L (i-1)} is the sum of all previous lower flank sums since the last cumulative sum is reset, and Y _i is the current laboratory measurement of the controlled output variable. The value according to the previous embodiment is therefore I ₂ No.
_LAB or DBP _LAB , GOAL is the target value of the output variable to be controlled, so according to the previous embodiment it is I ₂ No. _GOAL or DBP
Can be _GOAL , and k is the allowed slack in the controlled output variables (sl
ack), which is usually within a standard deviation of about 1 or each controlled output variable (I ₂ No.
About 68% of the values measured in the laboratory of _LAB or DBP _LAB ) fall into that.

The decision interval (-h, h) is set for each controlled output variable, and its exact value is empirically selected in the particular carbon black reactor used, but it is usually set for that output variable. It is near the allowable limit. For example, a typical value of h for iodine value or DBP can be 5. Therefore, the determination interval h would be 5 iodine value units or DBP units on either side of the value of I ₂ No. _GOAL or DBP _GOAL , respectively.

After taking each sample of carbon black and determining the value of the iodine value (I ₂ No. _LAB ) and / or the value of DBP (DBP _LAB ) measured in the laboratory, these values are about (Y _i ). Substitute into equations (30) and (31). Then 2
The two cumulative sums S _{H (i)} and S _{L (i)} are I ₂ No. _LAB and DBP
Calculate for both _LABs . Then the iodine value or D
If SH _(i) ≥h or SL _(i) ≤-h for BP, an alarm signal is generated for each output variable.
When an alarm signal occurs, the operator is instructed to increase the frequency of sampling of the carbon black produced, usually at least doubling it. The Kalman increase for the iodine number algorithm (K _I ) and / or the DBP Kalman filter increase for the algorithm of DBP (K _D ) respectively when the alarm signal occurs for the iodine number and / or DBP, respectively. Each set equal to 0. After taking a sample of the next carbon black, I ₂ No. _LAB or DBP _LAB is I ₂ No.
If it falls within ± k of _GOAL or DBP _GOAL , the cumulative sum is reset for each change by setting the cumulative sum _{SH (i-1)} and _{SL (i-1)} to zero. However, if the alarm signal continues to occur, increase the Kalman filter for each output variable (K _I or
K _D ) is set equal to 1 until the laboratory measured value falls within ± k of the target value for that variable.

In FIG. 2 the hardware components of the process control system of the present invention are shown schematically. The process control system consists of a system controller, generally designated 10. The system controller 10 is of the type known in the art, and is preferably a minicomputer, such as the VAX minicomputer described above. The system controller 10 is coupled to a control system 14 distributed through a bus 12. The distributed control system 14 is also of the type known in the art,
For example, the Fisher PRoVOX Instrumentation System (Fisher Conrols Internationa
, Inc., Marshalltown, Iowa). The distributed control system 14 sequentially processes the PID algorithm (P
ID) through an oil flow meter 16 and an automatically adjustable flow valve 18. As mentioned above, the oil flow meter 16 is preferably a Coriolis type flow meter. An oil flow valve 18 is installed upstream or downstream from the oil flow meter 16 in the carbon black reactor feedstock line 20. Therefore, the distributed control system 14 controls the actuation of the flow valve 18 to achieve a target iodine number (I ₂ No. _GOAL ), as further detailed below.
To achieve. The distributed control system 14 is also coupled to a potassium addition solution flow meter 22 and an automatically adjustable flow valve 24 through the PID algorithm (PID). The flow meter 22 is preferably a Coriolis type flow meter similar to the oil flow meter 16. Flow valve 24 is located upstream or downstream from flow meter 22 in line 24 of the potassium addition solution of the carbon black reactor. Therefore, the distributed control system 14 also controls the actuation of valve 22 to automatically adjust the flow rate (K ⁺ S) of the potassium addition solution to target DBP, as further detailed below. Achieve the value (DBP _GOAL ).

Referring to FIG. 3, there is shown a flow chart that conceptually describes the procedure of the carbon black process control system of the present invention. The indications S _{1 to} S ₁₂ indicate steps 1 to 12. When the process control system is implemented as shown in S ₁ , the distributed control system 14 has a predicted iodine number (I ₂ No. _P ) according to the iodine number algorithm and the DBP algorithm, respectively, as described above. And the predicted DBP value (DBP _P ) occurs as shown in S ₂ . Preferably, the iodine value algorithm, and thus the equation for the forward feed input variable, is embodied as a subroutine in the distributed control system 14. Similarly, the equations of the DBP algorithm are also preferably implemented by the distributed control system 14 in subroutines. After the I ₂ No. _P and DBP _P have been calculated, they are then each stored in the computer memory at the system controller 10. The distributed control system 14 calculates both I ₂ No. _P and DBP _P approximately once per second based on the current input variable reading, as shown at S ₃ . Each updated I ₂ No. _P and DBP _P is then saved in memory at system controller 10. Then, as shown in S ₄ , the values of I ₂ No. _P and DBP _P stored in the memory of the computer over each two-month interval were averaged by the distributed control system 14 to obtain I ₂ No. _AVG and DBP. Save it in _ANG and computer memory.

Based on the I ₂ No. _AVG over a 2 month interval, the flow rate of the new feedstock (OIL _NEW ) is then determined by the distributed control system 14, as shown at S ₅ . Similarly,
The flow rate (K ⁺ S _AVG ) of the fresh potassium addition solution is also determined based on the DBP _AVG over 2 months. As mentioned above, equations (5) to (7) and equations (17) to (2
1) is preferably embodied as a subroutine in a distributed control system 14 for determining both the fresh feedstock flow rate (OIL _NEW ) and the fresh potassium addition solution flow rate (K ⁺ S _NEW ) respectively. . Based on the flow rate of the new feedstock (OIL _NEW ) and the flow rate of the new potassium-added solution (K ⁺ S _NEW ), the control system 14 then distributed, as described in further detail below, PI
Valve 18 and valve 2 by using the D algorithm
Determine the degree to which 4 is adjusted. Each of the fresh feedstock flow rate (OIL _NEW ) and the fresh potassium addition solution flow rate (K ⁺ S _NEW ) is then updated every two months. Then, valve 18 and valve 24, in turn, each receive a new I ₂
Adjusted based on No. _AVG and DBP _AVG to achieve the new flow rate, as shown at S ₇ .

First in the offline laboratory measurement features of the present invention
Step is shown in S _8, which is the system controller 10, (at intervals spaced or other interval) per second during the period of taking a sample of carbon black is calculated I ₂ No. _P
Calculate the mean and standard deviation of both and DBP _P. The carbon black produced is sampled at spaced intervals, eg, typically within about 1 to 4 hours, and
Both iodine value and DBP of the sample are measured in the laboratory as shown in S ₉ (I ₂ No. _LAB and DBP _LAB ). As previously mentioned, the carbon black sampling interval is in the range of about 2 to 20 minutes. Then the new system intercept (KO) about the iodine value algorithm
_Was calculated by the system controller 10 during the sampling period, as shown at S ₁₀ , and I ₂ No. _LAB and
Update based on I ₂ No. _AVG . Preferably, equations (8)-(14) are embodied as a subroutine in system controller 10, as described above. Similarly, the scale factor (F) is also adjusted based on DBP _LAB and DBP _AVG during the sampling period.
Preferably, equations (22)-(29) are implemented as a subroutine in distributed control system 14, as described above. The new system intercept (KO _NEW ) was then used to determine a more accurate predicted iodine number (I ₂ No. _P ) until the next carbon black sample was taken, as shown in S ₁₁ . To update the algorithm for iodine value. Similarly, using a new scale factor (F _NEW ), update the DBP algorithm to determine a more accurate DBP value until the next carbon black sample is taken, as shown in S _11. . As shown in S ₁₂ , the iodine number algorithm and the DBP algorithm are each updated whenever a carbon black sample is taken, and thus within about every 1 to 4 hours.

Referring to FIG. 4, the iodine value (I ₂ No. _P ) according to the iodine value algorithm and the DBP _P according to the DBP algorithm
A flow chart is shown that conceptually describes the procedure of the distributed control system 14 in predicting both of the above. Distributed control system 14, first, as shown in S _1, the input variables of the forward feed algorithm for iodine value and
Read the input data needed to calculate the input variables for the DBP algorithm. Input variables for the iodine number algorithm include feedstock flow rate, gas flow rate, air flow rate, air preheat temperature, and air humidity. As mentioned above, ATBG (fuel quality) is a calculated control variable and ATBO (feedstock quality) is an essentially constant control variable. The input variables for the DBP algorithm are the potassium addition solution flow rate and the feedstock flow rate.

After reading the input data, the distributed control system then compares the input data to a range of acceptable values for each variable, as shown at S ₂ . If the value is outside its acceptable range (BAD), a bad data flag, which is a digital signal, is set, as shown at S ₃ . When the bad data flag is set, I ₂ No. _P and / or DBP _P are not calculated based on that data. If all the data are within the acceptable range, then both I ₂ No. _P and DBP _P , as shown in S ₄ , by using the iodine number algorithm and the DBP algorithm, respectively. Calculated based on input data. Both I ₂ No. _P and DBP _P are then compared to the actual range in which each output variable will fall, as shown at S ₅ . If I ₂ No. _P or DBP _P is not within the allowable range,
It sets a bad data flag, and the current values for I ₂ No. _P and / or DBP _P are not used, depending on whether one or both are outside their respective tolerances. If I ₂ No. _P or DBP _P fall within their tolerances, their values are stored in the computer memory of the system controller 10, as shown at S ₆ , and later (at intervals of intervals). At the end)
Each is used to update the feedstock flow rate and the potassium addition solution flow rate, respectively.

Referring to FIG. 5, there is shown a flow chart that conceptually describes the procedure of the distributed control system 14 for adjusting both the feedstock flow rate and the potassium addition solution flow rate. If the bad data flag is set as shown in S ₁ , and during the iodine value and / or the DBP prediction procedure (BAD) as shown in S ₃ in FIG. 4, the bad data flag is cleared. And the adjustment procedure shown in FIG. 5 is not performed during that interval as to which algorithm had bad input data. However, if the bad data flag is not set during the 2 minute interval, then the distributed control system 14 will set a new feedstock setpoint (OIL _NEW ) and / or potassium spiked solution setpoint, as shown at S _2. Read the input data to determine (K ⁺ S _NEW ).
The input data for OIL _NEW are AIR _AVG , GAS _AVG , ATBG and OAC _AVG as defined in equation (7).
Includes. The input data for K ⁺ S _NEW is K ⁺ S _AVG , OIL _AVG , DB, as defined in equation (21).
_Includes P _AVG and X _AVG .

Then, the input data, as shown in S _3, compared with the allowable range of values for each term. If the values fall outside their tolerances, the Bad Data flag is set (BA
D). Therefore, the feedstock flow velocity set point (O
IL _NEW ) and potassium addition solution set point (K ⁺ S _NEW )
Is not adjusted for its spaced spacing if one and / or both are bad. If all the values fall within their allowable ranges, as shown in S _4, OIL _NEW and
Each K ⁺ S _NEW is updated as described above. Then OI
Both L _NEW and K ⁺ S _NEW are compared to acceptable values, as shown in S ₅ . If OIL _NEW or K ⁺ S _NEW are not within their respective tolerances (BAD), the end of each term and its flow rate procedure is not adjusted. If OIL _NEW and K ⁺ S _NEW are within their tolerances, each of OIL _NEW and K ⁺ S _NEW is processed by the PID algorithm and S
Update the feedstock flow rate and the potassium addition solution flow rate, respectively, as shown in ₆ .

Referring to FIG. 6, a typical algorithm, preferably used to adjust the flow rate of new feedstock (OIL _NEW ) or the flow rate of new potassium addition solution (K ⁺ S _NEW ), is It is shown schematically. Each of the feedstock flow meter 16 and the potassium addition solution flow meter 22 are respectively coupled to a flow transmitter (FT). Each flow transmitter (FT), in turn, is coupled to a distributed control system 14 and transmits a signal (F _m ) corresponding to the flow velocity sensed and measured by its respective flow meter. Each of the signals for the new flow rate set points for the feedstock and potassium spiked solution (F _SP ) is then compared to the respective measured flow rate signal (F _m ) generated by the flow meter.
Based on each comparison, each flow velocity setpoint signal (F _SP ) -each measured flow velocity signal (F _m ).
An error signal (e (t)) equal to is generated for each respective flow velocity. Then each error signal (e
Known in the art based on (t)),
Each DBP algorithm produces an output signal (c (t)) corresponding to the adjustments to be made to each flow valve 18 or 24 to achieve the flow rate set point. The output signal is then sent to the respective current to the air converter (I / P). Each of the currents to the air converter (I / P) is coupled to an oil flow valve 18 and a potassium addition solution flow valve 24, respectively, to regulate each respective valve. Thus, each of the currents to the air converter (I / P) produces a pressurized output corresponding to its respective PID output signal (c (t)), which in turn produces a PID output signal (c (t)). The respective valves are adjusted to achieve the flow rate set point. Therefore, each PID algorithm has an error signal (e
The output signal (c (t)) continues to be generated until (t)) is no longer present and thus the flow rate set point is reached.

Referring to FIG. 7, at the end of each carbon black sample period, the system controller for updating the system intercept (KO) of the iodine number algorithm and / or the scale factor (F) of the DBP algorithm 10 A flow chart conceptually describing the procedure of is shown. As shown at S ₁ , the system controller recalls from memory the I ₂ No. _P and DBP _P calculated and stored during the sampling period.
If the system controller cannot properly recall the data (success), the algorithm will not be adjusted. Next, System Controller 10 Current I ₂ No. _LAB
And read the values for DBP _LAB and compare them with the values in the acceptable range. If either value is out of range, its respective algorithm is not adjusted. The system controller 10 then uses the cumulative sum procedure, which is the current I ₂ N, as shown at S _3.
o. Current sum _{SH (i)} and / or for _LAB and DBP _LAB values
Or determine S _{L (i)} . S _{H (i)} ≧ h or S for any measured output variable (I ₂ No. _LAB or DBP _LAB ).
_{If L (i)} ≤-h, the system controller generates an alarm signal. If an alarm signal is generated,
The increase of the Kalman filter for the algorithm of iodine number (K _I ) and / or the increase of the Kalman filter of DBP for the algorithm of DBP (K _D ) depends on whether the alarm signal is generated for one or both output variables. Dependently, set equal to 1. Therefore, the new system intercept for the iodine number algorithm (KO _NEW ) and / or the new scale factor for the DBP algorithm (F _NEW )
Both are based solely on laboratory measured values for I ₂ No. _LAB and DBP _LAB , respectively. However, if no alarm signal is generated, the system controller will use the new filtered analytical properties, I ₂ No. _FILTER and DBP.
_{The FILTER} is determined and then the intercept constant (KO) and scale factor (F) of the system are adjusted to update the iodine value algorithm and the DBP algorithm, respectively, as shown at S ₄ . Then, as shown in S ₅ , the value for the intercept of the new system (KO _NEW ) and the value for the factor of the scale (F
_NEW ) Compare each value with the allowable range. If either value is out of range, it is not used to update the respective algorithm. If each of the values for KO _NEW and F _NEW are within range, they are, as shown in S ₆ ,
Saved in memory. When the value is stored in memory, the system controller 10, as shown in S _7, then until the end of the sample period, clears the data of the inlet flag.

[Brief description of drawings]

FIG. 1 can use the control system of the present invention. 1 schematically shows an example of a furnace carbon black reactor. FIG. 2 schematically illustrates the hardware components of the process control system of the present invention. FIG. 3 is a flow chart conceptually showing the procedure of the process control system of the present invention for controlling the iodine value and / or DBP. FIG. 4 is a flow chart conceptually showing the procedure of the process control system of FIG. 2 when predicting the iodine value and DBP according to the present invention. FIG. 5 illustrates the distributed control system procedure of FIG. 2 for adjusting the feedstock flow rate and the potassium addition solution flow rate to achieve a target iodine number and a target DBP, respectively, in accordance with the present invention. It is a flowchart which shows typically. Figure 6 shows the target iodine value and the target DBP, respectively.
1 schematically illustrates the PID algorithm used in accordance with the present invention when adjusting the feedstock flow rate and the potassium addition solution flow rate to achieve FIG. 7 schematically illustrates the procedure of the system controller of FIG. 2 to adjust the iodine number algorithm and the DBP algorithm at the end of each carbon black sample period. 10 ... System controller, 12 ... Bath, 14 ... Distributed control system, 16 ... Oil flow meter, 18 ... Flow valve, 20 ... Feedstock line, 22 ... Potass solution flow meter, 24 ... Automatic adjustment can flow valve, 26 ... potassium additive solution line, S ₁ to S ₁₂ ... step 1-12.

Continuation of front page (72) Inventor Melvin Sea. Dennis USA, Texas 79065, Pampa, Pee. Oh. Box 5001, 3 Miles West Highway 60, Pampare and Day, Cabot Corporation (72) Inventor David J. Cowl USA, Texas 79065, Pampa, Pee. Oh. Box 5001, 3 Miles West Highway 60, Pampar & Day, Cabot Corporation (72) Inventor James L. Rice USA, Georgia 30328, Atlanta, Peachtree Dunwoody Road 6600, Aviation Law, Building 300, Suit 500, Cabot Corporation (72) Inventor Thomas El. Weber USA, Massachusetts 01821, Billerica, Concord Road 157, Billerica Technical Center, Cabot Corporation

Claims (5)

[Claims] 1. The following steps: measuring at least one input variable utilized in the production of carbon black, at intervals, while the carbon black reactor is operating, and using at least one algorithm. And predicting at least one carbon black output variable at spaced-apart predicted intervals based on the at least one input variable measured during the spaced-apart intervals; An average value of at least one predicted output is determined over the averaging interval and at spaced intervals during which the average value of the at least one predicted output variable and the reactor are operating. Adjusting the at least one input variable according to an adjustment algorithm using the difference between the target value of the at least one output variable, Achieving a target value for its output variable to obtain a substantially constant amount of carbon black, comprising controlling carbon black production in a carbon black reactor. 2. The following steps further comprising: sampling carbon black produced during the operation of the carbon black reactor at spaced sampling intervals, and Measuring the at least one output variable predicted by the algorithm from a sample of carbon black and adjusting the at least one algorithm based on the measured value of the at least one output variable to obtain the at least one output A method for controlling the production of carbon black according to claim 1, comprising the more accurate prediction of variables. 3. The method further comprising the step of: monitoring the measured value of the output variable of the at least one carbon black to detect an undesired shift in the average of the at least one output variable. 2. A method for controlling the production of carbon black according to 2. 4. A component: a metering means for measuring at least one input variable utilized in the production of carbon black at spaced intervals during operation of the carbon black reactor, coupled to said metering means. Calculating means for predicting at least one output variable of carbon black at spaced intervals according to at least one algorithm using said at least one input variable measured during said spaced intervals; The calculating means further determines an average value of the at least one predicted output variable at spaced average intervals, and is coupled to the calculating means for calculating the at least one predicted output variable. The difference between the mean value and the target value of the at least one output variable is used to A carbon black reactor comprising: adjusting means for adjusting the at least one input variable at spaced intervals to obtain a substantially constant amount of carbon black according to an adjustment algorithm for achieving a desired value of the force variable. For controlling the production of carbon black in. 5. Further constituents: carbon black produced during operation of the carbon black reactor is sampled at spaced intervals,
Sampling means thus enabling the at least one output variable to be measured from a sample of carbon black, wherein the calculating means is responsive to the measured value of the at least one output variable in response to the at least one output variable. 5. An apparatus for controlling carbon black production in a carbon black reactor according to claim 4, comprising adjusting said at least one algorithm by utilizing said measurement of a variable to more accurately predict said output variable. .