Real time optimization of MIP technology

catalytic cracking in Qingdao Petrochemical

 

Abstract: Real time optimization of self adaptation which is implemented on the fluid catalytic cracking unit of the MIP-CGP technology by using the dynamic correlation integration optimizer is introduced in this paper. Such situations as optimization targets, optimized variables and application effects are introduced. The industrial application results of the optimizer on 1.4Mt/a heavy oil catalytic cracking (RFCC) unit showed that: the system operation was stable, related parameters of the reaction regeneration tended to stable and the unit could adapt to change of properties of raw materials automatically; both the yield of liquefied gas and the total liquid yield were increased, in which the yield of liquefied gas was increased by 2.31% and the total liquid yield was increased by 1.07%; the changes of dry gas yield and coke forming ratio were not obvious and the economic benefits were increased by about USD 1/barrel. Therefore very good optimization effects were obtained.

 

1 Introduction

There are rather high difficulties in real time optimization of the fluid catalytic cracking unit. The challenge is how to trace the optimum operation point automatically under the situations of variable raw material properties, modification of unit and updating of catalysts, but no manmade interference is needed, such as reconstruction of models, modification of parameters. For an application example introduced in this paper, the dynamic correlation integration technology is adopted so that such a self adaptation optimization control can be realized.
The MIP-CGP technology of Research Institute of Petrochemical Processing is adopted for the fluid catalytic cracking unit of heavy oil in this example. The design processing ability of the unit is 1.4Mt/a.
In 2007, aiming at the MIP-CGP technology, the dynamic correlation online optimization software of Beijing OptimiPro Control Technology Co., Ltd. is adopted to implement online optimization on the unit. The real time optimization control system was successfully put into effect on March 30th 2007. Since the system was put into effect, the system is stable and operation is in good condition. Stable control and optimized operations on the unit, especially automatic adjustment of main operation conditions of the unit according to current raw materials and prices of finished products in the market made distribution of the products to comply with the principle of maximum benefits, which played positive roles.
The newest control technology is adopted for the real time optimization of the unit, so that the system can run under the conditions of variable raw materials, updating of catalysts, modification of the unit and change of system characteristics, and the system can find the main operation conditions which are the most suitable for current operation conditions, which makes the set objective function to be maximum. The system can realize the following objectives:

  1. Long term, safe, stable and reliable closed ring optimization control.
  2. Taking the yield of individual product as the objective function to optimize the key production conditions, so as to make the yield of objective product to reach the maximum value.
  3. Optimized production of multi processing schemes. The system can implement optimization operation on processing schemes whose main products are liquid hydrocarbon, gasoline, diesel oil respectively. The system can also optimize the total liquid yield of the whole unit.
  4. Control on product distribution. Under the possible conditions, the system can not only meet the requirements of product distribution, but also optimize the yield of the main product.
  5. Optimization of the overall economic benefits. The system can adjust automatically the main operation variables according to the market prices of raw materials and finished products so as to make the economic benefits obtained by the unit to process every barrel of raw materials to reach the optimum value.
  6. The catalyst average activity of the automatic online optimization unit which can reach the optimum matching with the current processing scheme.
  7. Self adaptation to immeasurable factors, such as change of the properties of processing raw materials.
  8. Self adaptation to operation status, such as changes of the processing amount of the unit, residual doping ratio and catalysts.
  9. Permission of optimization on parts of variables. That is to say, computer optimization control is realized for a part of key variables while the other variables are set manually.
  10. Adoption of automatic fault tolerant optimization control technology. The system can implement fault detection and positioning and determines that which optimized variables should be offline and online under the current status through an automatic decision-maker. The system implements optimization in permitted sub spaces.
  11. An emergency system. There is an emergency system under abruptly happening emergencies so as to ensure safe production.
  12. Constraint control – optimization coordination advanced control. It is used for harmonizing the Constraints in the system (such as the limiting boundary on level of the reprocessing oil tank) and the optimization control system.

2 Composition of the system

 

The system is a secondary computer control and it consists of one group of optimization control and an advanced control sub system. The system includes:

  1. Catalytic cracking optimization control system whose main product is liquid hydrocarbon
  2. Catalytic cracking optimization control system whose main product is gasoline
  3. Catalytic cracking optimization control system whose main product is diesel oil
  4. Catalytic cracking optimization control system whose target is the total liquid yield
  5. Catalytic cracking optimization control system whose optimization target is the economic benefits of the unit
  6. Optimization control scheme switching system
  7. Automatic fault detecting and fault tolerant optimization control system
  8. Level of the refining oil tank – optimization coordination advanced control system
  9. Level of the fractioning tower - – optimization coordination advanced control system
  10. Reaction – regenerating catalysts average activity optimization control system
  11. Emergency processing system
  12. Real time accumulated yield and benefit calculating system
  13. Operator and engineer interface
  14. Online soft measurement of dry point of gasoline and freezing point of diesel oil
  15. Automatic feeding system of IV-B type FCC new catalysts

 

2.1 Optimization control schemes

 

According to the present market situations of oil product selling and variability of the catalytic cracking processing schemes, the whole optimization system can adapt to different processing materials and product different main products. The optimization control system can adapt to five processing schemes:

  1. For optimization scheme whose main product is liquid hydrocarbon, the yield of diesel oil should not be smaller than one set value.
  2. For optimization scheme whose main product is gasoline, the yield of diesel oil should not be smaller than one set value.
  3. For optimization scheme whose main product is diesel oil, the yield of liquid hydrocarbon should not be smaller than one set value.
  4. Optimization scheme whose target is the total liquid yield.
  5. Optimization scheme whose target is the economic benefits of the overall unit.

 

2.2 Objective function

 

In the optimization scheme whose main product is liquid hydrocarbon, the yield of liquid hydrocarbon is taken as the objective function (OBJ)
OBJ= output of liquid hydrocarbon in 12 hours /processing amount of the unit in 12 hours
Constraint conditions: the yield of diesel oil should not be smaller than 25% (the value can be set manually)

In the optimization scheme whose main product is gasoline, the yield of gasoline is taken as the objective function (OBJ)
OBJ= output of gasoline in 12 hours /processing amount of the unit in 12 hours
Constraint conditions: the yield of diesel oil should not be smaller than 18% (the value can be set manually)

In the optimization scheme whose main product is diesel oil, the yield of diesel oil is taken as the objective function (OBJ)
OBJ= output of diesel oil in 12 hours /processing amount of the unit in 12 hours
Where: flow quantity of oil materials feeding into the unit = flow quantity of wax oils + flow quantity of oils in the first section + flow quantity of vacuum residual oils
Constraint conditions: the yield of liquid hydrocarbon should not be smaller than 10% (the value can be set manually)

In the optimization scheme whose target is the total liquid yield, the yield of liquid hydrocarbon + gasoline + diesel oil in 12 hours is taken as the objective function (OBJ)
OBJ= (output of liquid hydrocarbon in 12 hours + output of gasoline in 12 hours + output of gasoline oil in 12 hours ) / processing amount of the unit in 12 hours

In the optimization scheme whose target is the total economic benefits of the unit, the total economic benefits of the unit is taken as the objective function (OBJ)
OBJ = (output of dry gases in 12 hours × unit price of dry gases
+ output of liquid hydrocarbons in 12 hours × unit price of liquid hydrocarbons
+ output of gasoline in 12 hours × unit price of gasoline
+ output of diesel oil in 12 hours × unit price of diesel oil
+ output of discharged oil slurry in 12 hours × unit price of oil slurry
- feeding amount of wax oils in 12 hours × unit price of wax oils
- feeding amount of residual oils in 12 hours × unit price of residual oils
- addition amount of new catalysts in 12 hours × unit price of new catalysts) / processing amount of the unit in 12 hours – other unit processing costs

The prices of raw materials and finished products in the above equations can be modified at any moment in the environment of engineer.

These four processing schemes can be switched at any moment according to the instruction of the technological control.

2.3 Optimized variables

 

In selection of optimized variables, those variables which are important for the reaction-regeneration technology and are easy to be controlled were mainly considered. In the fluid catalytic cracking unit, the following variables are taken as the online optimized variables:

 

  1. Reaction temperature of the first reactor
  2. Reserve of catalysts in the second reactor
  3. Pre-lifting steam (steam pre-lifting technology)
  4. Pre-lifting dry gas (dry gas pre-lifting technology)
  5. Residual doping ratio
  6. Heat exchanging temperature of the raw material oil
  7. Flow quantity of the termination agent
  8. Reprocessing oil ratio
  9. Oil slurry reprocessing ratio
  10. Addition amount of new catalysts

 

2.4 Constraint conditions

2.4.1 Constraint of the optimized variables
In the optimization control system, in order to ensure safety and normal operation of the processes, all optimized variables can only change in one permitted range. Such kind of permitted range is the constraint of optimized variables. In the system, both upper limit and lower limit are set for every one optimized variables. Considering that constraints on these variables will change with the operation status of the unit and optimization schemes, these variables can be set manually through the operator interface according to demands.

 

2.4.2 Constraints of product distribution
In optimization control, optimization on main product might cause decrease of yield of other products. For example, in the diesel oil optimization scheme, increasing the yield of diesel oil will decrease the yield of liquid hydrocarbons. While in practical production process, sometimes the following requirement will be put forward: the yield of diesel oil should be increased in so far as possible under the condition that the yield of liquid hydrocarbon will not be smaller than one value. This kind of requirement can be regarded as a Constraint optimization.

In the optimization system, the constraint conditions on product distribution are added and they are shown in table 4.

Table 2.1: Constraints on product distribution in various optimization schemes


Optimization scheme

Constraint conditions on product distribution

Liquid hydrocarbon

The yield of diesel oil should not be smaller than one manually set value

Gasoline

The yield of diesel oil should not be smaller than one manually set value

Diesel oil

The yield of liquid hydrocarbon should not be smaller than one manually set value

The manually set value of Constraint conditions in table 4 can be modified by engineers or the constraint conditions can be removed.

 

2.5 Optimization methods and optimization software

The dynamic correlation integration optimization technology is adopted for optimization control of the unit. The dynamic correlation integration optimization technology has been applied on the dewaxing process and fluid catalytic cracking unit of the refining plant. In addition, the technology experienced change of processing oil products, maintenance and updating of equipments, and it has been subjected to the practice test of long term stable operation.

The dynamic correlation integration optimization technology can be applied to most of continuous production processes. Practices showed that this method has the following features:

1) No dynamic or static model of the system should be established in advance. Due to the complexity of the fluid catalytic cracking process, it is very difficult to establish a precise model for online optimization. Moreover, once the components in the unit are replaced or large changes happen on processing raw materials and catalysts, the adopted process model must be adjusted or updated again, while adoption of the dynamic correlation integration method can solve the problem easily.

2) The system runs by using the natural pulse in the production and no manual test signal is needed. No effect or interference will be produced on normal operations of the process during optimization.

3) The system can overcome the dynamic interferences whose characteristics are unknown existing in the optimized variables and objective functions. For example, the fluctuation of the objective function caused by the changes of properties of catalytic cracking processing raw materials will not generate large effects on the optimization process.

4) Strong adaptability. The theory and practices showed that the dynamic correlation integration optimizer had very strong adaptability on the characteristics changes of the optimized system. For example, when the process properties change due to modification of equipments, the system can work well. Practical applications showed such an advantage of the technology sufficiently.

The product of Beijing OptimiPro Control Technology Co., Ltd is adopted as the optimization software of the system: Dynamic Correlation Integration Optimizer Express for FCC for Centum CS3000 (FCC dynamic correlation integration optimizer Express Centum CS3000).

2.6 Level of the reprocessing oil tank – optimization coordination advanced control system

In order to ensure that the level of the fractioning tower bottom changes within the normal range, the level – optimization coordination advanced control is added for these two variables such as reprocessing ratio and addition amount of new catalysts which will affect the level of the reprocessing oil tank directly. The controller harmonizes the level of the fractioning tower bottom and level of the refluxing oil tank according to the rule as well as optimizes these two systems, which will keep the level of the fractioning tower bottom and level of the refluxing oil tank between 30% ~ 70% in so far as possible during the course of optimization control.

2.7 Fault tolerant optimization and emergency system

An abnormality detecting system of the measuring signals is provided for the optimization control. The system checks the input signals once every minute. If abnormal situations of the signals are found, an alarming signal will be given on the operator interface to alert the operator to perform examination on corresponding instruments. At the same time, the core of optimization control will reconstruct the optimization control according to the detection results of the abnormal detecting system. The system will suspend the optimization on those variables which are affected by the abnormal measuring signals and keep the status before abnormality, while continues optimizing those variables which will not be affected by the abnormal measuring signals. When the abnormality or measuring signals detected by the abnormality detecting system disappears, the core of the optimization control recovers the reconstruction automatically so as to restore the normal control pattern.

When emergent situations occur on the unit, the operator can start the emergency system. After the red button is pressed and confirmation is made, the emergency system should exclude all variables from the optimization control. At the same time, the system should switch the basic controllers related to the optimization control from remote setting to manual setting. The control right of all circuits should be returned to the operator.

When the unit recovers to the normal state, the operator can put variables into optimization according to normal procedures.

2.8 Reaction – regeneration catalyst average activity optimization control system

During the course of reaction, aiming at different requirements on product distribution, there are different optimum catalyst average activities. The system adjusts the average activity of system catalysts through controlling addition amount of new catalysts. While addition amount of new catalysts is implemented through a set of IV-B type automatic small feeding unit. An independent optimization controller calculates the optimum feeding amount through certain methods and performs the pulse feeding of new catalysts through an automatic small feeding unit which is controlled by DCS.

2.9 IV-B type FCC new catalyst automatic feeding system

The IV-B type FCC new catalyst automatic feeding system which obtain the national patent of China (ZL 90 1 07310.5) is a pulse automatic feeder for new catalysts in the fluid catalytic cracking unit which is controlled by DCS. The feeder feeds materials in pulse pattern. Compared to other automatic feeders, the feeder has such features as simple structure, high operation reliability and long life. The optimum feeding of new catalysts can be realized when the feeder is cooperated with the optimization control system, which plays important roles for adjusting and controlling activity of catalysts and reaching the optimum product distribution.

2.10 Control structure

The overall control structure of the system is shown in figure 1 and it is a secondary computer control pattern when seen in logic. Being the client, the optimization controller obtains the needed measurement information from the OPC Server connected to the DCS. The controller calculates the correlation integration between various variables and correlation integration between various variables and the objective function once every minute. The optimizer calculates the setting value once every ten minutes according to these correlation integration values and these setting values are sent to the basic controller for implementation.

dcio2

As such, the online closed ring optimization control is finished.

2.11 Human machine interface

The human machine interface consists of the operator interface and the engineer interface.
The operator interface mainly implements display of online detection data, alarming and normal operations.
The interface is mainly one main operation menu which displays the measurement values, gradient values, optimization values of various optimized variables, output values of the optimization controller, upper limit of the optimized variables, lower limit of the optimized variables, and alarming of the optimized variables as well as various states of the system.
In operation part, the workers can realize the following online real time operations through the keyboard or mouse: manual – automatic switching of any one or several optimized variables at any moment, modification of manually set values under the manual state, interference on automatic setting values under the automatic state.
The engineer interface can permit the engineer to input of modify necessary data of the system, such as various Constraint conditions, prices of finished products and raw materials, switching of the optimization schemes.

3 Application effects

3.1 Tests on effects of the optimization control system effects

In order to examine the effects of the optimization control system, one test is organized on the system. During the optimization test, the optimization scheme is the economic benefits of the unit. The variables of the optimization operation are the temperature of the first reactor, reserve of the second reactor, flow quantity of the pre-lifting steams, feeding temperature, feeding flow quantity of new catalysts. Other variables are kept constant and they will be not optimized. The results are shown in figure 3.1, table 3.1, table 3.2 and table 3.3.

 

dcio4
Figure 3.1: Comparison on economic benefits before and after optimization

It can be seen from figure 3.1 that the economic benefits of the unit appears the tendency of increase after the optimization control is put into effect. The detail change is shown in table 3.1.
Table 3.1: Change of economic benefits before and after optimization


Before optimization
(USD/barrel)

After optimization
(USD/barrel)

Increase amount
(USD/barrel)

11.47

12.56

1.09

Note: the economic benefit value in the table is the average value of economic benefits in one period before and after optimization.

 

Table 3.2: Change of the measurement value of the optimized variables before and after optimization


Optimized variables

Value before optimization

Value after optimization

Change amount (after optimization – before optimization)

Temperature of the first reactor

509.58

513.02

3.44

Flow quantity of steam in the pre-lifting

2.9977

2.65

-2.3477

Feeding temperature

187.23

183.04

-4.19

Feeding flow quantity of new catalysts

0.22887

0.16

-0.06887

Reserve of the second reactor

6.6558

7.04

0.3842

 

Table 3.3: Change of yield of various kinds of products after optimization is put into effect


Product

Before optimization

After optimization

Change amount (after optimization – before optimization)

Dry gas

4.514761

4.5800

0.0652

Liquid hydrocarbon

22.53039

22.6487

0.1183

Gasoline

35.46597

36.5448

1.0788

Diesel oil

21.30648

20.8828

-0.42368

Oil slurry

4.1216

4.3973

0.2757

Total liquid yield

79.3169

80.0766

0.7597

Note: the detail yield values in the table are the average value of accumulated yield of various products in one period before and after optimization.

 

It can be seen from the above test results that: after the optimization control is put into effect, the temperature of the first reactor is improved, the reserve of the second reactor is increased, the flow quantity of pre-lifting steams is reduced and the feeding temperature is decreased. Change of these variables can increase the reaction depth while decrease of the addition amount of catalysts can decrease the costs. Change of the total economic benefits is improved a little. Considering that the economic benefit optimization scheme embodies the comprehensive performances of the unit and the scheme has more practical application value when compared to other optimization schemes, therefore the economic benefit optimization scheme is mainly put into effect in the unit.
3.2 Long term operation situations and discussions
After being put into effect, the system attracts wide attentions of all aspects due to its excellent performances and good benefits. From the time of being put into effect to present, the system is kept in effect except for short time exclusion due to mechanical faults. The operation environments of the unit are modified greatly, the stability is increased and the benefits are improved.
The author compares the statistic material balance for three months before and after optimization respectively and the results are shown in table 3.4:

Table 3.4: Statistics on change of yields of various products before and after optimization is put into effect

Product

Before optimization

After optimization

Change amount (after optimization – before optimization)

Dry gas

3.71

3.89

0.17

Liquid hydrocarbon

22.21

24.52

2.31

Including: propylene

6.72

7.34

0.62

Gasoline

36.40

36.78

0.38

Diesel oil

22.08

20.46

-1.62

Oil slurry

6.67

5.52

-1.15

Singeing

8.50

8.45

-0.05

Loss

0.43

0.38

0.05

Total liquid yield

80.69

81.76

1.07

Conversion

71.25

74.02

2.77

It can be seen from the above table that the results of long term operation of the optimization system is consistent with the conclusions obtained from the test reports in the early stage of operation. After long term and constant optimization and operation examination, the effects are obvious and economic benefits are increased.

After the system is put into effect, the coke forming ratio of the reaction is reduced obviously and the total liquid yield is increased increasingly. Due to decrease of the coke forming ratio in the reaction, the control on temperature of the dilute and dense phases in the regenerator is more stable and rational.
Addition of catalysts is uniform, the activity is stable and the unit consumption is decreased.
The flow quantity of pre-lifting steams is adjusted automatically by the control valve according to the requirements of the reaction optimization so that the concentration of propylene in the liquid hydrocarbon can be increased.
Because the coke forming ratio of the reaction is decreased, the reaction depth can be increased in further, the yield of the gasoline and liquefied gas is increased obviously and the yield of light diesel oil is decreased obviously, which reaches the objective of producing more propylene in catalytic cracking.