[PPT] - Integration of Nonlinear CDU Models in Refinery CDU Models in PowerPoint Presentation

SLIDE 1

Integration of Nonlinear CDU Models in Refinery CDU Models in Refinery Planning Optimization

Abdulrahman Alattas, Advisor Ignacio Grossmann Advisor Ignacio Grossmann

Chemical Engineering Department Carnegie Mellon University

EWO Meeting – March 2011

1

SLIDE 2

I t d ti Introduction

 Refinery production planning models

 Optimizing refinery operation

C d l ti

 Crude selection

 Maximizing profit; minimizing cost  LP-based, linear process unit equations

, p q

 Current Project

 Collaboration with BP Refining Technology  Develop a refinery planning model with nonlinear

process unit equations

 integrate scheduling elements

2

g g

SLIDE 3

Refinery Planning Model Refinery Planning Model Development

3

SLIDE 4

LP Refinery Planning Model LP Refinery Planning Model Example

E ample

 Example

Complex refinery

config ration

Fixed yield Swing cut C d F d t k Crude1 (lighter) 142

configuration

 Processing 2 crude

ils & importing

Crude Feedstock Crude2 (heavier) 289 469 Other Feedstock Heavy Naphtha 13 9 Fuel Gas 13 17

ils & importing

heavy naphtha

Swing cut model

Refinery Production LPG 18 20 Light Naphtha 6 6 Premium Gasoline 20 20

 Offers lower net cost

& different feed quantities

Reg. Gasoline

80 92 Gas Oil 163 170 Fuel Oil 148 160 Net Cost 89663 85714

quantities

 Shows benefits of

better equations

Net Cost 89663 85714

4

SLIDE 5

Refinery Planning Model Refinery Planning Model Development

 Focus on the front end of the refinery  Focus on the front end of the refinery

Crude distillation unit (CDU)

5

SLIDE 6

CDU & C d d C l CDU & Cascaded Columns

Cascaded Columns Representation

f a Crude Distillation Column

(Gadalla et al, 2003) Typical Crude Distillation Column (Gadalla et al, 2003)

6

SLIDE 7

A t M d l Aggregate Model

V D Vtop Ltop Top D 1 Feed F Vtopfee

d

Ltopfeed p Section Feed Top F VtopFeed LtopFeed Vbotfee

d

Lbotfee

d

Bottom Section Bottom VbotFeed LbotFeed B Vbot Lbot B Steam n

7

Conventional distillation Steam distillation

SLIDE 8

NLP R fi Pl i M d l NLP Refinery Planning Models



Aggregate Model



Mixed-type distillation cascade

Combines conventional and steam distillation

 Combines conventional and steam distillation  Challenges with full CDU model  4+ cascaded columns

32

Feed

 32+ components

Bottom Section Feed F Bottom Section 8

SLIDE 9

FI M d l I t FI Model - Intro

Dist4

CDU is a series of fractionation

units

TC3 TC4 Dist4 Dist3 Dist2

 Cut point temperature is the

separation temperature

TC1 TC2

C3

Feed Prod2 Prod3 Prod4 Dist1 Prod1

3 4 5 6 Component Distribution of A Distillation Column Using FI

B d G dd ’

1

1 2 3

1.5
1
0.5

0.5 1 1.5 2 2.5 3 Log (XDist/XProd)i

Slope=1.95

Based on Geddes’

fractionation index method (Geddes 1958)

9

4
3
2

1 Log io

Slope=3.40

method (Geddes 1958)

SLIDE 10

FI M d l E ti FI Model - Equations

 Mass balance  Temperature

Tcj  TEj TIj1

 Component

V

j

2

Tc j 1  Tc j

p distribution

 Vapor pressure

Pv  Pc * Exp 5.96346 1 Trj, i

 1.17639 1 Trj, i  

1.5

0.559607 1 Trj, i

 

3 1.319 1 Trj, i

 

6

       /Trj, i         10 Pvj, i  Pci * Exp i 4.78522 1 Trj, i

  0.413999 1 Trj, i  

1.5

0.891239 1 Trj, i

 

3  4.98662 1 Trj, i

 

6

       /Trj, i          

SLIDE 11

FI M d l R k FI Model - Remarks

FI Model is crude independent

 FI values are characteristic of the column  FI values are readily calculated and updated from

refinery data

A id l li d li

Avoids more complex, nonlinear modeling

equations

Generates cut point temperature settings for Generates cut point temperature settings for

the CDU

Adds few additional equations to the planning Adds few additional equations to the planning

model

11

SLIDE 12

FI M d l CDU E l FI Model – CDU Example

 FI model

GO

 FI model

 FI model example

 Venezuelan crude

TC TC3 TC4

HN Dist4 Dist3 Dist2 Dist1

TC5

N

 40 Pseudo-components, 5 cuts  4 cases:

 Maximizing naphtha (N), heavy naphtha (HN), light

TC

1

TC

2

Feed BR HD LD HN Dist1

g p ( ) y p ( ) g distillate (LD), heavy distillate (HD)

 Cut-point temperature and product quantities

reflect the different business objectives

 Stats  Equations: 562  Variables: 568

S l CONOPT

Product Run Gas OH Naphtha H NaphthaL Dist. H Dist. B. Residue Max Naphtha 6.2 112.9 35.1 68.6 16.5 60.7 M H N h 6 2 10 4 3 0 6 1 16 6 60 Cut point temperature Run Naphtha H Naphtha L Dist. H Dist.

B. Residue

Max Naphtha 272.7 417.0 426.4 526.8 595.3 M H N h 272 7 386 2 487 8 526 8 595 3

 Solver: CONOPT  Time: 0.360 sec

Max H Naph. 6.2 107.4 53.0 56.1 16.6 60.7 Max L Dist. 6.2 111.5 10.7 95.0 16.0 60.5 Max H Dist. 6.2 111.5 10.7 94.0 16.9 60.5

13

Max H Naph. 272.7 386.2 487.8 526.8 595.3 Max L Dist. 272.7 386.2 398.3 606.0 631.1 Max H Dist. 272.7 386.2 398.3 526.8 650.5

SLIDE 13

Pl i M d l E l Planning Model Example

Typical Refinery Configuration (Adapted from Aronofsky, 1978)

butane

Fuel gas Prem.

SR Fuel gas

Cat Ref

Crude1, …

Gasoline Reg. Gasoline

SR Naphtha SR Gasoline

Distillate blending

CDU

Distillate

SR Distillate

Product Blending blending Gas oil

Cat Crack

Crude2, ….

Fuel Oil

SR GO

Hydrotreatment

blending Treated Residuum

SR Residuum

14

SLIDE 14

Planning Model Example Planning Model Example Problem Statement

 Information Given

 Refinery configuration: Process units  Feedstock & Final Product  Feedstock & Final Product

Crude1 Louisiana Sweet Lightest Crude2 Texas Sweet Crude3 Louisiana Sour

 Cases: Processing 2,3 & 4 crude oils

Crude4 Texas Sour Heaviest Case 1 Crude1 Crude2

 Objective

Case 2 Crude1 Crude2 Crude3 Case 3 Crude1 Crude2 Crude3 Crude4

 Select crude oils and quantities to process

 Maximize profit  single period time horizon 15

SLIDE 15

Pl i M d l E l R lt Planning Model Example Results

Comparison with the fixed yield and swing cut

models E i

Economics

 FI calculates the maximum profit scenario

Model Case1 Case2 Case3 FI 245 249 247 SC 195 195 191 FY 51 62 59

16

SLIDE 16

Planning Model Example Results Planning Model Example Results

(cont.)

F d t k lt

 Feedstock results:

 Different crude purchase option

Model Crude1 Crude2 Crude3 Crude4

C d Oil F d C t ib ti f C 2

Model Crude1 Crude2 Crude3 Crude4 Model Crude1 Crude2 Crude3 Crude4 Case 1 FY 54 46 SC 90 10 FI 72 28 FY 10 41 49

70% 80% 90% 100% eed

Crude Oil Feed Contributions for Case2

Model Crude1 Crude2 Crude3 Crude4 Case 1 FY 54 46 SC 90 10 FI 72 28 FY 10 41 49 Case 2 FY 10 41 49 SC 80 10 10 FI 10 30 60 C 3 FY 10 31 49 10 SC 70 10 10 10

10% 20% 30% 40% 50% 60% Crude Oil Fe Crude3 Crude2 Crude1

Case 2 FY 10 41 49 SC 80 10 10 FI 10 30 60 C 3 FY 10 31 49 10 SC 70 10 10 10 Case 3 SC 70 10 10 10 FI 10 19 61 10

0% 10% FY SC FI Model Type

Case 3 SC 70 10 10 10 FI 10 19 61 10

17

SLIDE 17

Planning Model Example Planning Model Example Results (cont.)

P d t

 Products

 Increased reg. gasoline  Different fuel oil rates and treated residue Cases Product FY SC FI Case 1 Fuel Gas 12.7 9.5 16.3 Premium Gasoline 20.0 20.0 20.0 Regular Gasoline 20 4 23 4 39 0 Cases Product FY SC FI Case 1 Fuel Gas 12.7 9.5 16.3 Premium Gasoline 20.0 20.0 20.0 Regular Gasoline 20 4 23 4 39 0

90% 100%

Refinery Products Slate for Case2

Case 1 Regular Gasoline 20.4 23.4 39.0 Fuel Oil 29.5 48.4 25.7 HT Residuum 19.0 Fuel Gas 12.4 9.5 15.9 Premium Gasoline 20 0 20 0 20 0 Case 1 Regular Gasoline 20.4 23.4 39.0 Fuel Oil 29.5 48.4 25.7 HT Residuum 19.0 Fuel Gas 12.4 9.5 15.9 Premium Gasoline 20 0 20 0 20 0

50% 60% 70% 80% 90% Products HTR

Case 2 Premium Gasoline 20.0 20.0 20.0 Regular Gasoline 20.1 23.4 39.0 Fuel Oil 32.1 48.5 26.2 HT Residuum 17.0 Fuel Gas 12.3 8.4 15.8 Case 2 Premium Gasoline 20.0 20.0 20.0 Regular Gasoline 20.1 23.4 39.0 Fuel Oil 32.1 48.5 26.2 HT Residuum 17.0 Fuel Gas 12.3 8.4 15.8

20% 30% 40% 50% Refinery P FO RG PG FG

18

Case 3 Fuel Gas 12.3 8.4 15.8 Premium Gasoline 20.0 20.0 20.0 Regular Gasoline 19.9 20.3 38.7 Fuel Oil 32.1 52.0 26.6 HT Residuum 17.4 Case 3 Fuel Gas 12.3 8.4 15.8 Premium Gasoline 20.0 20.0 20.0 Regular Gasoline 19.9 20.3 38.7 Fuel Oil 32.1 52.0 26.6 HT Residuum 17.4

0% 10% FY SC FI Model Type

SLIDE 18

Planning Model Example Planning Model Example Results (cont.)

 Model statistics

 FI model increased the number of equations and  FI model increased the number of equations and

variables

 ~30% nonlinear variables  Solution time is short  Solution time is short Model Variables Equations Nonlinear Variables CPU Time Solver FY 14 130 6.7 CPLEX Case 1 CPLEX SC 163 140 7.3 FI 1225 1204 348 7.3 CONOPT Case 2 FY 185 161 8.4 CPLEX SC 215 176 9.0 FI 1808 1772 522 8 8 CONOPT

19

FI 1808 1772 522 8.8 CONOPT Case 3 FY 231 194 9.8 CPLEX SC 271 214 10.3 FI 2395 2342 696 10.8 CONOPT

SLIDE 19

C l i Conclusion

Proposed FI model Proposed FI model

 Crude independent  Calculates cut point temperature settings

S f l i CDU l l ti

 Successful in CDU calculations  FI-based planning model calculates higher profits

using different crude oil purchase decision

 Limited increase in solution time  Limited increase in solution time

Aggregate model

 Successful models for Conventional distillation and

steam distillation steam distillation

 Resolving modeling full CDU with mixed type

columns

 NLP models  NLP models

 Assess the benefit of the different modeling approaches in

terms of accuracy, robustness & simplicity

20

SLIDE 20

Thank You Thank You

21