You are on page 1of 127

ESSENTIALS OF REFINERY PROCESSES

S K KALRA INDIAN OIL CORPORATION LTD PANIPAT REFINERY

REFINERY

WHAT IS A REFINERY ??

PETROLEUM REFINERY
Crude Oil Marketable Products

REFINERY OVER VIEW


Corporate Refinery Management Information Services

OM&S

Maintenance/In spection (Civil,


Mechanical. Electrical, Metallurgy)

PRODUCTION

Process Section/Qual ity Control Engineering Services

Administration & Finance

Stores & Materials

COMPLETE REFINERY
Quality Control

Marketing
OM&S (Product Blending & Dispatch)

Power & Utilities

Finishing Units Separation Units Conversion Units

Crude Oil

Jetty / Pipeline

OM&S

OM&S
(Intermediates)

Waste Treatment

Welfare / Admn.

Around the clock operation

Complexity of a Refinery
The combination of refining processes and operations employed (complexity) varies from one refinery to another. Factors deciding the complexity of a refinery Nature/source of crude oils to be processed Demand pattern in the markets to be covered Product quality current / future Production of feed stocks for downstream units Inter-fuel substitution Environmental stipulations

Crude Oil Characterization

By Gravity:
oAPI

SG <0.85 0.85-0.8984 0.8984-1.00 >1.00 <0.5 0.5-1.0 >1.0

Light Medium Heavy Extra Heavy Sweet Medium sour Sour

>35 26-35 10-26 <10

By Sulphur (%wt.):

Crude Oil Characterization

4 types of crude oils available to refiners around the world:


oAPI

Sulphur (% wt.) <0.5 0.5-1.5 1.5-3 >3

Light Sweet Light Sour Heavy Sour Extra Heavy

30-40 30-40 15-30 <15

High Acid Crudes (HACs)


TAN (Total Acid Number) > 0.5mg KOH/gm Crude Oil

PROFITABLE CRUDE

TYPE OF CRUDE TO BE USED

Crude Availability Crude Cost Desired Product Yield Refinery Complexity Environmental Constraints

Depends on :
-

CRUDE OIL SELECTION & OPTIMISATION IN REFINERIES

Configuration of Refinery what are the units present Metallurgy of refinery particularly columns, piping to take care of acidic / corrosive crude Product demand in the region Netback/ GRM of a particular crude Availability of a particular crude at economic cost.

Crude Oil Selection


Overall refinery economics depend on Crude cost + Processing Cost Lower the S, lower the SG Higher is the crude price Lower processing requirement HACs are normally cheaper Higher neutralization cost Refineries would like to handle crudes with TAN<0.5 & subsequent process streams containing TAN<1.5

Crude Oil Selection


Crude availability shifting from Light sweet Heavy sour Extra Heavy HACs are opportunity crudes New Refinery: design to process Extra Heavy Crudes design to process HACs in admixture

MAJOR REFINERY PRODUCTS

LPG (Propane/Butane)/Propylene MS/Naphtha/Solvents/Benzene/Toluene ATF/SKO/MTO/LABFS HSD/LDO FO/ LSHS /HPS/CBFS/PROCESS OILS Asphalts/Bitumen Lube Oil Base Stocks( GR-I & GR 2) RPC / CPC Slack Wax/ MCW/Paraffin Wax

Critical Quality Parameters of Products


Sl.no
1.

Product
LPG

Key Quality parameters


Evaporation Temperature at 95 % Volume =2 deg C, Max Cu Corrosion =not worse than No.1 RVP =1050 KPa, Max Density RON Sulphur Benzene =720-775 Kg/M3 =91 Min =150 ppm, Max =1 Vol. %, Max

2.

Motor Spirit

3.

ATF

Density =775-840 Kg/M3 Flash Point = 38 deg c, Min Sulphur =0.25 wt %,Max Smoke Point =20 mm, Min Density =790-820 Kg/M3 Flash Point =35 deg c, Min Sulphur = 0.25 wt %, Max Smoke point = 18 mm, Min

4.

SKO

Critical Quality Parameters of Products

Sl.no
5.

Product
Diesel Density Sulphur

Key Quality parameters


=820-845 Kg/M3 = 350 ppm = 51 Min =95 Min =180 , Max (Summer) Sulphur Ash = 4 wt %, Max = 0.1 wt %, Max =60 (1/10mm) , Min =175 deg C,Min =45-55 deg C

Cetane Number Recovery at 360 deg C 6. Fuel Oil

Kinematic Viscosity@ 50 deg c =125 , Max (Winter)

7.

Bitumen

Penetration at 25 deg c Flash Point Softening Point

BUSINESS PROCESS FLOW

Crude evaluation & Procurement ?


BE/MOU/St Tar: Term contract crudes. ILP : Spot procurement for next 3 months.

What and How to Feed ?


BE/MOU/St Tar ILP

What & Where To Make ?


BE/MOU/St Tar ILP

Demand Forecast ?
By Mktg.

Distribution Planning ?
By Mktg.

What to Store and where ?


By Mktg.

BASIC REFINERY OPERATIONS

Hydrocarbon molecules in crude do NOT meet customer needs SEPARATION PROCESSES (Primary Processes) CONVERSION PROCESSES (Secondary Processes) FINISHING PROCESSES (Secondary Processes) Segregate the molecules

Rearrange the molecules

Remove Contaminants

MARKETABLE PRODUCTS

REFINERY SCHEME
Crude Pretreatment Desalting Fractionation Atmospheric & Vacuum Distillation
Conversion

Treatment Diesel Hydrodesulphurization Diesel Hydrotreatment Solvent Extraction ( e.g. FEU) Catalytic Reforming Formulation & Blending Other Refinery Processes a. Sour Water Stripping b. Sulphur Recovery

Fluidized-bed Catalytic Cracking Hydrocracking Delayed Coking Visbreaking /Coking

c. Cooling Water Treatment

20

10

11

Refinery Operations

PRIMARY PROCESSING UNITS SECONDARY PROCESSING UNITS

CONFIGURATION OF REFINERIES / REFINING PROCESSES


PRIMARY UNITS CRUDE DISTILLATION UNIT ( CDU )/ VACUUM DISTILLATION UNIT(VDU)

SECONDARY UNITS FLUID CATALYTIC CRACKING UNIT ( FCCU ) , HYDRO-CRACKING UNIT ( HCU) , DELAYED COKER UNIT ( DCU ) , VISBREAKER UNIT ( VBU) LUBE/WAX FURFURAL EXTRACTION UNIT ( FEU) / NMP EXTRACTION UNIT, PRODUCING UNITS SOLVENT DEWAXING UNIT ( SDU) , CATALYTIC ISO-DEWAXING UNIT ( CIDW ) , WAX HYDROTREATING UNIT ( WHU) , HYDRO-FINISHING UNIT ( HFU) TREATING UNITS CATALYTIC REFORMING UNIT ( CRU) DIESEL HYRDO-TREATING UNIT ( DHDT ), DIESEL HYDRO-DESULFURISATION UNIT ( DHDS ), MEROX UNIT, ETC

12

PRIMARY PROCESSING UNIT


The purpose of Primary unit is to separate the crude in to different fractions by distillation. Known as mother unit of the refinery, consist of CRUDE DISTILLATION UNIT (CDU) VACUUM DISTILLATION UNIT (VDU) Commonly referred as Atmospheric and Vacuum Distillation unit (AVU)

Heavy at the bottom, light on the top The separation of crude oil by atmospheric and vacuum distillation into groups of hydrocarbon compounds of different boiling point ranges (called fractions or cuts) The first step in crude oil processing The process unit where the first separation takes place is called Crude Distillation Unit (CDU), Atmospheric Unit (AU) or Atmospheric & Vacuum Unit (AVU) This step is performed in all refineries : These units are called Mother Units Typical products from CDU are : Gas, LPG, naphtha, SKO/ATF, HSD and RCO. Vacuum Distillation of RCO produces VGO (or LOBS cuts) & VR All products need further treatment/processing.

Separation

13

14

SEPARATION PROCESSES

Crude Oil Fractions of crude boil at different temperatures Components are separated by distillation and drawn off as they condense

Atmospheric & Vacuum Distillation Unit Flow Diagram


LPG
Pr= 40-60 mm Hg a

S T A B

Pre-topping

Stabilized Naphtha
Light Naphtha

To Steam Ejector

T=200 Deg C

Unstabilised Naphtha

LVGO Heavy Naphtha

Vacuum

ATF

LDO HVGO

Atmospheric

Heavy Kero LGO HGO

Desalter
T=130-140 Deg C FURNACE Furnace T= 370 Deg C

Vac Slop

Crude Oil RCO

FURNACE Furnace

SR
T= 415 Deg C

15

16

Diesel Hydro-Desulphurisation Unit (DHDS)

Objective : To meet the EURO-II diesel quality requirement (<500 ppm S) Feed : Straight run diesel / FCC diesel component/ Coker and Visbreaker diesel components. Catalyst : Ni-Mo oxides

Diesel Hydro Desulphurisation Unit


Fuel gas H2 Recycle H2 Make Up
A b s or b er

Off. gas

Temp. 350 Deg C

Lean Amine
S

COL OVHD RECEIVER

High Pr. Separator Pr. 35-40 kg/cm2g Feed

Rich Amine

T R I P P E R

Wild naphtha

Steam

Furnace

Reactor

Desulphurized Diesel

17

DHDS Product Yields and Operating Conditions

1.

Typical Product Yields Sl.no.


1. Products Off Gas Wt% 1.36 End Users Refinery Fuel gas system after Amine Wash To Naphtha Pool after stabilisation To Euro II Diesel Pool

2. 3.

Wild Naphtha Diesel

1.04 97.1

2. Operating Conditions : Temperature range System Pressure : 320-380 DEG C : 30-40 kg/cm2(g)

Diesel Hydrotreating Unit (DHDT)


Objective : To meet the Euro III/IV diesel quality requirement ( 350/50 ppm S and Min. 51 Cetane No.) Feed : Straight run diesel / FCC diesel component/ Coker and Visbreaker diesel components. Catalyst : Ni-Mo oxides Chemical reactions: Desulphurisation and Denitrification
S N H2 H H2 H2S NH3

Aromatic compound + H2 --> Napthene Compound

18

Diesel Hydrotreator unit Flow Diagram


Fuel gas H2 Recycle H2 Make Up
A b s or b er

Off. gas

Temp. 350 Deg C

Lean Amine
S

COL OVHD RECEIVER

High Pr. Separator Pr. 100-105 kg/cm2g Feed

Rich Amine

T R I P P E R

Wild naphtha

Steam

Furnace

Reactor

Desulphurized & Hydro treated Diesel

DHDT Product Yields and Operating Conditions


1. Sl.no.
1.

Typical Product Yields


Products Off Gas Wt% 2.65 End Users Refinery Fuel gas system after Amine Wash To Naphtha Pool after stabilisation To Diesel Pool

2. 3.

Wild Naphtha Diesel

2.8 96.1

2. Operating Conditions : Temperature range System Pressure : 320-380 DEG C : 100-105 kg/cm2(g)

19

20

Catalytic Reforming Unit (CRU) Objective : To Upgrade the Naphtha to High Octane MS Component (Reformate) . Feed : 85-160 Deg C cut Naphtha / Visbreaker Naphtha Catalyst : Ni-Mo Oxides for NHTU Reactor Pt-Sn or Re for Reforming

21

Catalytic Reforming Unit


Main types of reformers are : 1. Semi-regenerative (SR) The reformer processes feedstock for a time and then shuts down for regeneration. 2. Cyclic Any reactor can be isolated for regeneration while the other reactors are in operation 3. Moving bed or CCR Catalyst is moved continuous through the reactors, withdrawn from the last reactor, regenerated in regeneration section and returned to the first reactor as fresh catalyst.

Continuous Catalytic Reforming Unit Flow Diagram


LSRN (85-160) VBN

Off Gas to FG system NHDT

H2 RICH GAS TO PSA

Off Gas to FG system

Pr. 17.0 kg/cm2g

COL OVHD RECEIVER

RGC

Pr. 17.0 kg/cm2g

T =-10 deg c

Recontacting drum SEPARATOR

S T A B I L I S E R

LPG

Regenerator T=520-540 Deg C

Reformate to MS Pool via RSU

22

CRU Product Yield and Operating Conditions


1. Typical Product Yields Sl.no.
1. 2. Products H2 Rich gas LPG Wt% 6.5-8.0 2.5-12 Quality 94% H2 gas S free LPG End Users PSA Unit to recover H2 To MS Pool After catalytic Reforming To MS POOL after Reformate Splitter

3.

Reformate

80-91

RON- 98, low S , High Bz.

2. Operating Conditions : Temperature range System Pressure : 490-540 DEG C : 2.0 - 30 kg/cm2(g)

Isomerisation Unit (PENEX-DIH)


Objective : To Upgrade the Naphtha by increasing its Octane Number to Higher Octane/Low Benzene/Low Olefins MS Component (Isomerate) to Meet Euro III / IV MS Specifications. Feed : C5-85 Deg C cut Naphtha /FCC gasoline(70-90 deg C cut)/ Lt. Reformate Catalyst : Co-Mo Pt Ni Based for Hydrotreator Reactor for Penex Reactor for Methanation

23

Isomerisation Unit Flow Diagram


LSRN (C5-85) FCC H/C (70-90) Hydrotreated Naphtha H2 Make up gas

Off Gas to FG system NHDT Lt. reformate


S T A B I L I S E R

Off Gas to FG system

Pr. 20.3 Kg/cm2g


LPG to Stripper Pr. 1.3 Kg/cm2g

COL OVHD RECEIVER

DIH Recycle

D I H C O L U M N

Penex feed dryer Meth nator reactor

Feed surge drum

Make up gas dryer

Penex Reactors

Temp. = 125-185 Deg C

Total Isomerate

ISOM Product Yield and Operating Conditions


1. Typical Product Yields Sl.no.
1. 2. 3. Products Off gas LPG Isomerate Wt% 1.4 11.3 87.3 End Users Refinery fuel gas System To LPG POOL To MS POOL

2. Operating Conditions : Temperature range System Pressure : 126- 145 DEG C : 33.5 kg/cm2(g)

24

Cracking process
Fluidized Catalytic Cracking Unit Hydro Cracker Unit Visbreaker Unit Coking unit

TECHNOLOGICAL ASPECTS

FCCU / RFCCU

Heavier Hydro-Carbon molecules are cracked under severe operating conditions of Temp. ( 500 510 oC ) and pressure ( 1.4 - 2.2 kg/cm2) to get Lighter Hydro-Carbons like LPG , MS & HSD components. Strict operating conditions are maintained to get on-specs. products. Heavier Hydro-Carbon molecules are mixed with Hydrogen and the mixture is subjected to severe operating conditions of Temp. (380 - 400 oC ) and pressure (165 185 kg/cm2 ) to get Lighter Hydro-Carbons like LPG , MS & HSD components. Strict operating conditions are maintained to get on-specs. products. All products are of Superior quality w.r.t. Sulfur content.

HCU / OHCU

25

Fluid Catalytic Cracking Unit (FCCU)

Objective : To convert Heavy Vacuum Gas Oil to valuable distillates like LPG, Gasoline, Diesel by catalytic cracking in fluidized bed. Feed : VGO/RCO/VR/HydroCracker Bottom. Catalyst : Silica & Alumina Zeolite Structure

26

YIELD PATTERN OF VARIOUS FCC UNITS


VGO FCC
FEED FEED QUALITY CCR, WT% S, WT% VR CONTENT, WT% PRODUCT, WT % GAS LPG PROPYLENE GASOLINE DIESEL (TCO) FO COKE PRODUCT KEY PROPERTIES GASOLINE : RON DIESEL CETANE INDEX 89 30.9 92.9 29.6 96 18 95 30 VGO

RFCC
VGO+VR

INDMAX
VGO+VR+EXTRACT

PETROFCC
VGO

0.74 3.4 NIL

4.06 3.59 20 MAX

<10 <44

0.74 3.4 NIL

1.47 8.68 3.04 20.06 52.64 7.98 5.00

3.00 8.79 3.71 18.60 46.45 10.82 7.43

10.44 16.80 11.20 28.00 10.10 8.60 13.80

8.80 21.00 22.00 28.00 9.50 5.00 5.50

Fluid Catalytic Cracking Flow Diagram


FLUE GAS M A I N C O L U M N COL OVHD RECEIVER HN FG LCO

S T A C K ORIFICE CHAMBER

Regen erator

R E A C T O R

HCO Recyc le

LPG WGC GAS CON UNIT

FCCG MAB

CLO
AIR CO BOILER

FEED
FURNACE Furnace

SLURRY SETTLER

Slurry Recycle

27

FCCU Product Qualities & End Users

Sl.No
1. 2. 3. 4. 5.

Product
Gas LPG Gasoline Hy.Naphtha + LCO CLO

Qualities
H2S rich Off. Gas H2S, Mercaptons, olefins like Propylene/Butylene

End Users
Refinery Fuel gas System after Amine Wash To LPG Pool/ Petrochemical feedstock

High Octane No. and high MS Pool Olefin contents Low Cetane no. High S , Unsaturates High Aromatics, Good Cutter Stock Diesel Pool After Hydrotreatment Fuel Oil

Hydrocracker Unit

Objective : To convert Heavy Vacuum gas oil to valuable distillates like LPG, Naphtha, ATF, Kerosene and Diesel. Feed : VGO / Coker Products Ni/Mo oxides for Dematalisation & Hydrotreating

Catalyst :

Ni/Mo/W(Tungsten) for Hydrocracking

28

Hydrocracking Process
Feedstock : VGO Products & Yields:

Gas LPG Naphtha SKO/ATF HSD Unconverted

2.5% 2.5% 8% 25% 22% 40%

Good process for increasing distillates and producing finished products. Existing at Gujarat, Mathura, and Panipat refineries.

HCU
FEED FEED QUALITY CCR, WT% MAX S, WT%, MAX N, PPM, MAX Ni+V, PPM, MAX SODIUM, PPM PRODUCTS, WT % GAS LPG NAPHTHA KEROSENE DIESEL BOTTOM VGO

OHCU
VGO

1 2.8 800 1.25 1

1 2.8 800 1.25 1

2.52 4.57 11.33 39.58 41.81 0.00

3.27 1.95 9.13 12.00 50.65 25.00

29

HCU
PRODUCT KEY PROPERTIES NAPHTHA: RON S, PPM, MAX KEROSENE : SMOKE POINT, MM FREEZINFG POINT, OC DIESEL : CETANE INDEX S, PPM POUT POINT, OC

OHCU

72 10

72 10

22-23 < 60

22-23 < 60

62 < 10 - 12

56 < 10 - 12

Various configurations of Hydrocraker Units

1. Single stage Once through Hydrocraker unit (SSOT): a. Feed and Hydrogen is passed through reactors only once for 60 80 % of partial conversion. b. Unconverted Oil is sent to FCCU. 2. Single stage recycle (SSRec) : a. Unconverted oil is recycled back to feed for 100% conversion.

3. Two stage Hydrocraker Unit: a. Unconverted Oil of SSOT is sent to another reactor for 100% conversion

30

Single stage Vs Two stage Hydrocraker unit


Makeup H2 Fresh Feed
R E A C T O R

Recycle Gas

Gases

Recycle Gas
RGC

Heavy Naphtha Kerosene Diesel

RGC

separator

First stage Product

Unconverted Oil

separator

Second stage Product

1) Single Stage in Blue 2) Two Stage is Blue and Green Combined

HCU Product Yields and Operating Conditions

1.

Typical Product Yields Sl.no.


1. 2. 3. 4. 5. 6. Products Off. Gas LPG Naphtha ATF/Kero Diesel Hydrocraker Bottom Wt% 2-4 1.5-3 6.5-10 27-40 29-40 5- 35

2. Operating Conditions : Temperature range System Pressure : 370-420 DEG C : 160-170 kg/cm2(g)

31

Hydro Cracker Unit Flow Diagram


Ex MPU
Makeup H2

RGC
MUC

R-1
PSA

R-2 Wash Water

HP Amine Amine Scrubber

Gas & LN to Light End Recovery Section

F-1 G-1 Feed Filter E-1


V-1

Heavy Naphtha ATF/Kero

CR V-2 V-3

Hydrocracking Reactor
PRT

MPU
To MUC V-8

Unconverted Oil

Diesel

AM HVGO

Feed Preheat Ex
BH HVGO

V-9
F-2

HCU Product Qualities & End Users

Sl.No
1. 2. 3. 4. 5. 6.

Product
Gas LPG Naphtha ATF / Kero Diesel Unconverted Oil

Qualities
H2S rich Off. Gas H2S Contents low Octane No. and low S contents Low S and Low Aromatics Low S and High Cetane Low S, High Saturates

End Users
Refinery Fuel gas System after Amine Wash To LPG Pool after caustic wash To Gasoline Pool / Hydrogen unit Feed To ATF/ kero. Pool EURO III Diesel FCCU FEED

32

Visbreaker Unit (VBU)


Objective : To reduce viscosity of Heavy Ends i.e. RCO/Vacuum residue by Thermal Cracking. Feed : RCO/Vacuum residue/Asphalts Typical Operating Conditions: Temperature Range Pressure

: 450-470 Deg C : 9-14 kg/cm2(g)

Viscosity of Feed : 500-3000 cst at 100 Deg C Viscosity of Product (VBtar) : 50 300 cst at 100 deg C

33

Visbreaker Unit Flow Diagram


T=135 0C P=4.6 Kg/cm2 FUEL GAS
V B N S T A B

COL OVHD RECEIVER

P=10 Kg/cm2 g

M A S I O N A C K O E L U R

VB HN G O S T R

S O A K E R
SLOP

VAP RET

VBN steam
GAS OIL

GO steam
VB TAR STR

M N

FEED

Furnace Furnace

T=450 0C

LP GAS TO FURNACE

VB FO

Visbreaking Process

Feedstock : Vacuum Residue Products & Yields:


Gas+loss Gas oil FO 3% 2% 93% Naphtha 2%

Good for FO production. Other products unstable and need further treatment Existing at Gujarat, Haldia, Mathura & Panipat refineries

34

VBU Product Yield/Qualities & End Users


Sl.No Product
1. Gas

Yield
1.82

Qualities
H2S rich Off. Gas

End Users
Refinery Fuel gas System after Amine Wash To FCCU or CRU

2.

VB Naphtha VB Gas Oil VB Tar

3.12

H2S, Mercaptons, high olefins Low Cetane no, Highly unsaturated Lower Viscosity than feed

3.

13.9

To DHDS or Fuel Oil

4.

81.16

Fuel Oil

Delayed Coker

Objective : To produce valuable distillate from Heavy ends by thermal cracking. Feed : RCO/Vacuum Residue/other heavy ends or residues Typical Operating Conditions: Temperature Range : 495-505 Deg C Pressure : 2-3 kg/cm2(g)

35

Delayed Coking Process

Feedstock : Vacuum Residue / VGO Products & Yields :

Good process for increasing distillates and minimising black oil production.Gas oil & Naphtha need further treatment. Existing at Barauni, Guwahati & Digboi refineries and under commissioning at Panipat

Gas+loss LPG Naphtha Gas oil FO RPC

10% 4% 5% 53% 11% 17%

Delayed Coker Unit Flow Diagram

Gas/LPG

FEED

F R S A O C A T KI O E N A R T O R

Naphtha Kero Gas Oil Fuel Oil

S E P A R A T O R

C O K E C H A M B E R

C O K E C H A M B E R

RFO

Pre heat Exchanger

Furnace Furnace

T=505 0C

36

Delayed Coker Product Yield/Qualities & End Users


Sl.No
1. 2. 3. 4. 5. 6. 7.

Product
Gas LPG Naphtha Kerosene Gas Oil Fuel Oil Coke

Qualities
H2S rich Off. Gas Mercaptons, unsaturates Low Octane, High Olefins High unsatuartes Low Cetane No. and high unsaturates Good cutter stock Low ash, High Sulphur

End Users
Refinery FG after Amine Wash To LPG after Merox /Caustic wash To FCC or CRU To DHDT Feed To DHDT & HCU feed Fuel Oil Gasification/Electrode Preparation/ cement ind.

Treating Process Caustic wash Merox Unit

37

Treating Process
1.

Caustic Washing for removing H2S and light Mercaptans and suitable for LPG and Naphtha Merox Process a. Extractive Merox : Suitable for Lighter fractions b. Sweetening Merox: Suitable for boiling range upto 350 Deg C

2.

Chemical Reaction: Amb.Temp. 2RSH + O2 Iron group metal


chelates

RSSR + H2O

After Treatment: The treated stream is given water wash, followed by sand Bed Coalescer or salt drier for removing entrained water.

Production of Lubricating Oil Base Stock

38

Lube Base Oil Processing

Crude Selection Multi-step manufacturing process

Lube Base Oils Key Properties:


Viscosity Viscosity Index Pour point Colour Flash point Volatility Oxidative & Thermal Stability

39

API Base Oil Characterization Groups


Viscosity Index GROUP I GROUP II GROUP III GROUP IV GROUP V 80-120 80-120 >120 Saturates % wt <90 >90 >90 Sulphur %wt >0.03 <0.03 <0.03

Poly Alpha Olefins(Synthetic Oils) All other base oils

Lube Base Oil Processing Philosophy


Properties controlled by Process Units Gas
KV, Flash Point

VI, CCR
Waxy Raffinate

Pour Point

Colour & Stability

Atmospheric distillation

Vacuum distillation

Hydrofinishing

Naphtha VGO Vacuum Distillates Kero Diesel

De-waxed H 2 oils

Crude oil

Solvent Extraction (Furfural & NMP)

Solvent Dewaxing Unit

Group I LOBS 150N 500N HN BN/150BS

VI,Pour Point, Colour&Stability

VR RCO

DAO

Propane DeAsphalting

H2

Catalytic Dewaxing Unit

100N 150N 500N Group II LOBS 150BS

40

Haldia Refinery
VGO (SO,LO,IO,HO)

LOBS Production Future Outlook


SOLVENT EXTRACTION (FURFURAL/ NMP) 150N SOLVENT DEWAXING + HYDRO FINISHING 500N HN BS API GROUP I BASE OILS

V A C U U M D I S T

HEAVY EXTRACT

VGO (SO,LO, IO,HO) DAO PROPANE DEASPHALTING

HYDROCRACKER

FUELS (65-75%)

BOTTOMS (25-35%)

VR

V A C U U M D I S T

70N VISCOSITY BASED CUTS CATALYTIC DEWAXING 150N 500N API GROUP II BASE OILS

RFCCU ASPHALT

BLOCK FLOW DIAGRAM REFINERY


CRU

GAS

SRU
MSQ

SULPHUR
LPG MS NAPHTHA

crude
RCO

CDU-I CDU-II
HU KHDS

MTO ATF/RTF SKO

DHDS

HSD JBO(P)/JBO(C)

VGO SO VDU-I VDU-II

FCCU

OHCU

MCW

MCW 100N 150N 500N

LO IO NMP HO PDA SR ASPHALT DAO

FEU

SDU

HFU

HLD-HN SL.WAX

150-BS ISO DEWAXING BITUMEN VBU


CBFS FO

IFO

41

API Gr I and Gr II Processing Schemes

Solvent Extraction

Solvent Dewaxing

Hydrofinishing

Solvent Extraction

Hydro Treating

Wax HydroIsomerisatio finishing n

Group I

Group II

Pour point VI

Other Process Units


Hydrogen Generation Unit Bitumen Blowing Unit Sulphur Recovery Unit

42

Hydrogen Generation Unit (HGU)


Objective : To Meet the Hydrogen requirement for DHDS/DHDT/OHCU/ISOM/Reforming Units and Other Hydrotreaters. Feed : Natural Gas / Naphtha Catalyst : Co-Mo ZnO/K2Co3 NiO Ni CuO Adsorbents for Hydrotreater for H2S and Chloride adsorber for Preformer for Reformer for HT/LT Shift reactors for PSA Adsorbers

HGU Product Purity and Operating Conditions


1. HGU Product is 99.99% Pure Hydrogen
2. Operating Conditions :

Temperature range System Pressure

: 860-870 Deg C : 23-26 kg/cm2(g)

43

Hydrogen Generation Unit Flow Diagram


FEED
R-01

Pre-Reformer R-03

HYDROGEN SAT. STEAM F-001 Reformer


Steam Drum R-04
DM water
De- aerator

R-02A/B

PSA

PSV HP STEAM Export FLUE GAS AIR condensate

OFF GAS as FUEL

Bitumen Blowing Unit


Objective : To Produce different grades of Bitumen by air blowing of vacuum residue at high temperature. Bitumen is colloidal solution of asphaltenes and high molecular gums in the medium formed by oils and low molecular gums. Feed : Vacuum Residue

44

BBU yield /quality and Operating Conditions


1. Typical Product Quality Sl.no.
1. 2. 3. Products Off gas Recovered liquid cut (FLO) Finished Bitumen Wt% 0.86 0.26 98.88

2. Typical Operating Conditions of Bitumen Blowing Unit: Temperature Range Pressure : 230-260 Deg C : 0.5 kg/cm2(g)

Bitumen Blowing Unit Flow Diagram


Air
Bitumen reactor
S E P A R A T O R

Off gasTo Incenerator

Furnace

FEED

Furnace

Bitumen to R/D after Cooling

FLO

45

Sulphur Recovery Unit


Objective : To Reduce the SO2 emission from the Refinery by recovering Sulphur from Amine Acid and Sour Gases produced during various Hydrotreating Process. Feed : Amine Acid gases and Sour acid gases

Sulphur Recovery Unit Flow Diagram


H2S/CO2 gas Rich amine A R U
AIR Claus Reactors 1/2/3 T=195-310 Deg C AIR Process gas (Superclaus reactor/TGTU) T= 220 deg C

Offgas
Off-gas incineration To stack

Lean amine

H2S / NH3 gas

Sulphur

T= 730 Deg C

Sour water

S W S U Stripped water

Sulphur degasification

Product sulphur handling

46

SRU Product Yield and Operating Conditions


1. Typical Product Yields Sl.no.
1. 2. Products Off gas Sulphur Wt% 0.1 99.9 End Users To Stack after incineration Sulphur Yard for Dispatch

Typical Operating Conditions: Temperature Range Pressure : 195-320 Deg C : 0.56 kg/cm2(g)

Other refinery Processes/operations

Steam & Power generation Process and DM water systems Hydrogen,nitrogen and air systems Flares and relief systems Sulfur recovery system Waste water treatment systems Safety & fire fighting systems Quality control, maintenance and administrative systems

47

REFINERY- UTILITIES

ENERGY REQUIREMENT

ENERGY CONSUMPTION
ELECTRICITY
FUEL GAS

FUEL
FUEL OIL

STEAM

LP

MP

HP

Energy consumption differs in different refineries due to: Refinery configuration / complexity Crude oil composition Technology / Equipments efficiency

48

ENVIRONMENTAL CONCERNS

SULPHUR in PRODUCTS BENZENE in PRODUCTS EFFLUENT WATER QUALITY SOx, NOx, H2S, Toxic Gases OTHER WASTE

MINIMIZING POLLUTION

Operate Furnaces Efficiently


Unrecovered Light Ends burnt in flare Stack

Avoid Spills & Accidental Releases Special Treatment of Sewer Water

49

MINIMIZING POLLUTION

FACILITIES TO BE IMPROVISED AT DESIGN STAGE


Adequate Stack height for better dispersion of pollutants Desulphurisation of fuel gas Provision of a Sulphur recovery unit Provision of continuous SO2 analyzers in all stacks Providing Air monitoring stations Efficiently running Effluent Treatment Plants New Unit / Up-to-date technology for producing Ultra low Sulphur and benzene free fuels.

REFINERY OPERATIONS HAZARDS

Low Flash Point Fouling and Corrosion Develop Static Charge


PRODUCT RELATED HAZARDS

Low Explosivity Limits Low self ignition temperatures Pyrophoric Toxic

Carcinogenic

50

TECHNOLOGY ADOPTION IN REFINIERIES

Desulfurization of fuel products for reduction in Sulfur- DHDS unit, Kero-HDS unit, DHDT Conversion processes for bottom of the barrel upgradation - FCC, Hydrocracker, DCU etc. Quality Improvement to meet environment norm- Cetane improvement in Diesel; Benzene, Olefin, Aromatics & Sulfur reduction in Motor Spirit. Adoption of catalytic-dewaxing technology for Quality Lube.

CHALLENGES TO OIL INDUSTRY

Environmental pressure- key factor in development & acquisition of new technology Sophistication in equipment design- demands for high performance products. Adoption of Euro norms for environment friendly transport fuels production, viz.,Gasoline & Diesel. Demand for environment friendly, high quality LOBS- API class-II/ III. Cost Intensive Refining Technology.

51

CONSTRAINTS TO MEET THE CHALLENGES

Crude oil sourcing - Indigenous production is only about 30% of the total requirement. Sharp fall in the availability of low Sulfur crude oil and even to the extent lighter crude oil.

Hence refineries are forced to process wide variety of crude oil including high sulfur crude. Selection of suitable technology having enough flexibility.

EMERGENCE OF COST INTENSIVE REFINING TECHNOLOGY

Switch over to Automation & Advance Controls Upgradation of the bottom of the barrel Efficiency Improvement thru debottlenecking / low cost revamps etc. Environment friendly processes for pollution abatement Stringent quality products manufacture & QC.

52

FUTURE CHALLENGES
TOTAL DEREGULATION COMPETITION FROM PRIVATE REFINING COMPANIES PRODUCT QUALITY STRINGENT MARKET DYNAMICS MARGIN PRESSURE CUSTOMER FOCUS

STRATEGIES

VALUE ADDITION CAPACITY SATURATION QUICK RESPONSE QUALITY / QUANTITY COST REDUCTION EFFECTIVE MANNING INTEGRATION FORWARD/BACKWARD/ LATERAL

53

Refining Vision

Refinery Capacity Refinery Margins Product Quality

ESSENTIALS OF REFINERY PROCESSESSES

INDIAN OIL CORPORATION LIMITED

Thank you

54

Phase Equilibria in Refinery Processes

Ratan Mohan

Department of Chemical Engineering I.I.T. Delhi, New Delhi

Thermodynamic data needs in process simulation

Phase equilibria

Stream properties ; enthalpy, entropy

Reaction equilibria ; Gibbs free energy of rxn, Eq. constt.

Basic Phase Equilibrium equation :

Hence,

Property Calculations :

Activity Coefficient method :

where

For non-condensing gaseous components :

Liquid property calculations :

Equations of State :

Cubic EOS :

Activity Coefficient Models :

Van Laar Scatchard-Hildebrand

Margules Redlich Kister

Wilson NRTL UNIQUAC

UNIFAC

Non Random Two Liquid (NRTL) Model:


applicable to partially miscible as well as completely miscible systems The NRTL equation for the excess Gibbs energy x j G ji ji E
g = xi RT i
j

x G
k k

ki

Activity coefficient in its generalized form is given by


(nG E / RT ) ln i = ni P ,T ,n j

ln i

G x xG = + G x G x
ji ji i j j ij ki k j kj k k

xk kj Gkj ij k Gkj xk k

contd
where: i, j, k = 1, 2,c ;

ij =

( g ji gii ) RT

G ji = exp ( ji ji ) ;

ij ' s & ij are NRTL model parameters


where

ij ji ij = ji = 0.2
for liquid liquid system

UNIFAC (Universal Functional Activity Coefficient) method


estimates activity coefficients based on the group contribution concept Excess Gibbs energy (and logarithm of the activity coefficient) as a combination of 2 effects1. 2.

combinatorial term residual term


ln iC = ln i x i z + qi ln i 2 i

ln i = ln i + ln i
C

where
i =
c

+ li i xi

NOG j =1

x l

j j

xi ri

x r
j =1

; i =

xi qi

j j

x q
j =1 j

; li =
j

z (ri qi ) (ri 1) 2

ri =

NOG k =1

i k

R k ; qi =

NOG k =1

v Q
i k

contd
ln iR =
where
NOG k

k i

(ln k ln ki )

k = residual activity coefficient of group k in the mixture ki = residual activity coefficient of group k in a reference solution containing only molecules of type i.

The parameters k and ki are defined by:


NOG NOG ln k = Qk 1 ln m mk NOGm km m m n nm n
c

k =

X k Qk
NOG m

mn = e

bmn / T

Qm

Xk is the group vkj x j j X k = c NOG mole fraction of group k in the liquid: vmj x j
j m

Property methods for Petroleum mixtures :

Eqn of State property methods for hydrocarbons at high pressure:

Peng-Robinson (PR) Most enhanced model in Aspen HYSYS Largest applicability range in terms of T and P Special treatments for some key components Largest binary interaction parameter database PRSV Modified PR model Better representation of vapor pressure of pure components and mixtures Extends applicability of the original PR model to moderately non-ideal systems SRK Modified RK model Can provide comparable results to PR in many cases, but with a lot less enhancement in Aspen HYSYS PR-Twu SRK-Twu Twu-Sim-Tassone (TST) Modified equations of state models for hydrocarbon systems-non ideal systems (used for glycol package) Generalized Cubic Equation of State (GCEOS) Provides a framework which allows users to define and implement their own generalized cubic equation of state including mixing rules and volume translation MBWR Modified BWR model Having 32 parameters, this model works extremely well with a number of pure components within specified T and P ranges Lee-Kesler-Plcker Also a modified BWR model for non-polar substances and mixtures BWRS Modified BWR to handle multi components Requires experimental data

Zudkevitch Joffee Modified RK model with better prediction of VLE for hydrocarbon systems, and systems containing hydrogen Kabadi-Danner Modified SRK model with the enhancement to improve the VLE calculations for H2O-hydrocarbon systems, particularly in dilute regions Sour PR/Sour SRK Used for sour water systems containing H2S, CO2, and NH3 at low to moderate pressures

Semi-empirical Models :
Chao-Seader model Applicable to hydrocarbon systems in the range of T=0-500C, and P<10,000 kPa Grayson-Streed model An extension to the Chao-Seader model with special emphasis on H2 Recommended for heavy hydrocarbon systems with high H2 content, such as hydrotreating units

Hydrocarbon systems up to distillate range hypo components PR, SRK or any other EOS*

Vacuum columns GS, PR or BK10

Sour gas sweetening with Amines

Sour water treatment process Sour PR/SRK

Clean fuels for sulfur components and hydrocarbons

High H2 content systems GS, PR

Utility systems using H2O Steam Table


Aspen

Refinery Processes :

Scope of Optimization in Refining Operation

The Presentation Structure

Basics of Optimization Optimization within Refinery Optimization among Refineries Petroleum Supply Chain Optimization

Basics of Optimization
A process to achieve best solution / performance within defined constraints

Profit Maximization
Throughput maximization within hardware constraints Maximize equipment life through optimum usage Optimization

Within Refinery

Among Refineries

Logistics Rail Pipelines

On-line

Off-line

Ref-1

Ref-2

Tankers Road

APC Optimizers

LP Models

Ref-3

Basics of Optimization
M N A . C U EPR E& C A K O TH VE R D IC R C S 180.00 160.00 140.00 120.00 $/BBL 100.00 80.00 60.00 40.00
JUNE ''08 OCT08 DEC08 A UG'08 M AR'08 NOV 08 A PR'08 FE '09 B JUL'08 S 08 EP

20.00
M Y A '08

M AR'09

A PR'09

JAN'09

G 0.5% O

M S

DB U AI

B ET RN

M Y A '09

Basics of Optimization
M N A . PR D C C A K O TH VE O U TS R C S 50.00 40.00 30.00 20.00 $/BBL 10.00 0.00 -10.00 -20.00
JUNE''08 OCT08 DEC08 A UG'08 JAN'09 FE B'09 JUL'08 S EP08

-30.00
M AR'08 APR'08 M AY'08

NOV08

M AR'09

APR'09
4 5

N HH AP T A

JE KE O T R

G 0.5% O

M S

Basics of Optimization
Optimization
Profit Maximization Throughput maximization within hardware constrainits Maximize equipment life through optimum usage

Ref. Profit

=Prod. Realization Input Cost Operating cost = Sum(Qi*Pi) - Sum(Ci*pi) Sum (Fi*Ui) Losses = f ( Type of crude, Process Configuration, Demand Pattern) => Under Control = f ( Demand -> domestic, International) => Little control = f ( Global demand (Premium/Disc), Location of Ref. & Crude source) => No control = f ( Production rate, Global demand, Political scenario) => Under control

Qi (Prod. qty) Pi (Prod. Price) pi (Crude Cost) Ci (Crude Type)

M AY'09

Basics of Optimization Optimization within Refinery

Optimization within Refinery


GAS LPG NAPHTHA CRU CRUDE BH / ASSAM / GUJARAT LS HS MS

Product Mix Optimization

Naphtha/ MS / Pet-chem
ATF / SKO ATF / SKO / HSD FO / Bitumen

CDU

ATF SKO DHDS HSD

FCCU VDU HCU Coker BBU VBU

MS

HSD MS / HSD / Coke BITUMEN FO / LSHS

Optimization within Refinery


Functional Objective Maximize Gross Refinery Margin Input cost reduction Maximize capacity utilization to reduce operating cost Maximize value added products Swing Operation in FCCU Optimum utilization of VBU / BBU Minimize low value products irrespective of demand Operational Efficiency improvement - Fuel & Loss, R&M cost, Quality give-away etc. Net Margin = Products sold * Transfer Price - Crude process * Crude cost at Refinery - Operating Cost

Optimization within Refinery


Margins Function of Demand pattern Products Prices Product pattern of refinery & Its flexibility / Optimization Refinery configuration Logistic Infrastructure availability Optimization LPG Vs Propylene (PRU) Naphtha Vs MS (CCRU, FCCU, Isom, Crude type) SKO Vs ATF SK Vs HSD FO / LSHS Vs HSD FO Vs Bitumen
9

Optimization within Refinery


On-line Optimization Advanced Process Control CDU Pressure minimization subject to constraints CDU COT maximization subject to constraints FCC Severity maximization / optimization Value added product maximization subject to property constraints Inferential properties prediction Constraint controller Multi-variable predictive control On-line Optimizer Offsite blend optimization

10

Optimization within Refinery- Offline


Product pattern Optimization Secondary units availability
FCC / RFCC HCU / OHCU Coker / VBU / BBU Naphtha cracker

Type of crude processing LS / HS / Hy. crude


Refinery operating cost

Flexibility of swinging product pattern Demand pattern


Seasonal demand (Naphtha, ATF, Bitumen) Prod. Price difference ( FO / HSD, LDO/ HSD, Naphtha vs Nat. Gas)

RTPs of products
11

Optimization within Refinery


Product Pattern Optimization Secondary Units- FCC / RFCC MS / LPG maximization Suitable for higher UOP K VGO feed Medium investment Medium Op. cost
CCR limitation Metal limitation Lower Cetane of TCO

Secondary Units- HCU / OHCU SK / ATF / HSD maximization without any treatment Lower UOP K VGO feed can be processed High investment High Op. cost
Nitrogen limitation
12

Optimization within Refinery


Product pattern Optimization Secondary Units- Coker Residue up-gradation Facilitates Hy. Crude processing Relatively low investment Relatively low Op. cost
Around 30% coke generation HPS vs Coke price deciding factor Cracked products needs treatment HCGO needs reprocessing in HCU / FCCU

13

Optimization within Refinery


Product Pattern Optimization Secondary Units- VBU Viscosity breaking to produce FO Lower cutter stock requirement Low investment Low Op. cost Dubai vs Brent, FO vs HSD deciding factor Secondary Units- BBU Suitable for VR having high Asphaltenes Releases cutter stock for value added products Seasonal and reagion specific demand No import facility Dubai vs Brent, FO vs HSD deciding factor
14

Optimization within Refinery


Input cost Optimization Input Cost = f ( Type of crude. Logistic cost, other inputs) Crude Type LS Crude ( Low S, High API, High Dist.)
Lower Operating cost Higher FOB and logistic cost

HS Crude ( High S, Medium API, Medium Dist.)


High Operating cost Medium FOB and Lower Logistic cost

Hy. Crude ( High / Low S, Low API, Low Dist.)


High Operating cost High viscosity affecting PL capacity High Acid no., Metal content Cheaper crude

The capability of type of crude processing will depend on Refinery configuration


15

Refinery LP Model
Model Crude Assays
Input Available Crude Basket Crude and Product Prices at refinery gate Product demand Shutdown Slate Change in product specs.
Refinery Production Planning Crude evaluation Blend Optimization Investment Analysis Inventory Management Back casting
16

Process unit configuration Feed Blends Product Blends Product Specs Objective : Profit Maximization

Output Crude Mix and Qty. Product pattern Sec Unit capacity utilisation. Sec Units feed and Product Blends

Provision for 24 Multi-periods

Basics of Optimization Optimization within Refinery Optimization among Refineries

17

Refineries configuration
U nits J C P.,MM A TPA 13.70 C U D VD U R C C FC /FC H U H U C /O C VBU BBU C KER O LO BS P/C E H MIC L A J; K Y LI, P:PA IPA O A N T,M :MA U A :B R U I, H A IA G U TH R ,B A A N :H LD , :G WA A D IG O H TI, ;D B I
18

IO L R FIN R S C E E IE P M B 12.00 8.00 6.00

H 6.00

G 1.00

D 0.65

A SSO IA S R FIN R S C TE E E IE B PL R C L PC C R B 2.35 9.50 1.00

Optimization Between Refineries


Synergy among multi-refinery operation to maximize over all profit Intermediate stream sharing Refinery configuration Planned shutdown schedule Capacity constraint in One refinery and availability in other Refinery SRGO Reformat PXN IFO (J-> P/M, BGR->G/B) (J/B/D -> G) ( M- > P ) (J- > P ) : HS crude & prod. Optimization : Optimization between Naphtha & MS : HS crude maximization : HS crude maximization : Naphtha export minimization

PNCP Feed (J/M - > P)

19

Basics of Optimization Optimization within Refinery Optimization among Refineries Petroleum Supply Chain Optimization

20

Corporate Supply Chain


Crude Imports
Domestic Domestic Crude Crude Imp. LS Imp. LS Crude Crude Imp. HS Imp. HS Crude Crude International Trade International Trade Crude Crude Products Products
Prod. Imports / Exports

Domestic Domestic Demand Demand

Prod. Transfer Refining Refining Mktg. Mktg.

Imports / Imports / Exports Exports OMCs OMCs exchange exchange

Support Up-Stream (High Lead time)

Pipe Lines / Pipe Lines /

Support Support Down-Stream

21

Indian Refineries
Capacity in MMTPA

IOC
BHATINDA (9.0)

Operates 10 of Indias 18 refineries refining Capacity: 60.2 MMTPA (1.2 mbpd) largest in the country 41% refining share in the country
DIGBO BONGAIGAON (0.7) NUMALIGARH (3.0) (2.4) BARAUNI GUWAHATI (1.0)

PANIPAT (6.0+6.0+3.0) MATHURA (8.0)

JAMNAGAR (33+30, 12) MUMBAI

BARODA (13.7)

BINA (6.0)

(6.0) HALDIA (6.0) PARADIP VISAKH (7.5+0.8) TATIPAKA (0.1) CHENNAI (9.5) (15.0)

(12, 5.5+2.4)

Total capacity in the country: 148.9 MMTPA (as of 1 Apr09)

MANGLORE (9.7)

Subsidiaries of IOC

COCHIN (7.5)

NARIMANAM (1.0)

Existing IOC New / + Additions 22

Others

Infrastructure
Crude Vadinar / Mundra port (VLCC) for North-West Refineries SMPL for Gujarat, Panipat & Mathura Mundra for Panipat Haldia Port for East coast Refineries Lower Draft and port congestion HBCPL for Barauni Refinery Commissioning of Paradeep- Haldia crude pipeline Product Demand growth in North West Sector Euro-III products & ATF demand in Metro cities Product movement from East to North-West Limited export facility at Haldia port Dahej / Kandla for Naphtha Export No Import / Export facility for ATF and Bitumen

23

Complexity in Indian Oil Supply Chain


40 Crudes in basket from S. America to S.E.Asia 10 Refineries Large distribution network

10 major products 200 Depots (excluding LPG network) 40 Terminals 17 Pipelines 4 Transportation modes
One crude pipeline catering to 3 refineries Crude procurement 3 months in advance

24

Down stream Oil Industry: Overview

Integrated Supply Chain Management


Plan

Source

Make

Deliver

Source

Make

Deliver

Source

Make

Deliver

Supplier
Up-stream time Mgmt M+4 Months

Your Organization
SCOR model
Source: The Supply Chain Council

Customer
Down-stream time Mgmt.

IOCL,BPCL,HPCL
25

Crude evaluation / purchase


Line up crude term contract for 50-60 % of requirement Spot / short term purchase for balance quantity Buy crude giving maximum Supply chain margin Based on landed crude price & domestic products pricing Crude purchase at Vadinar & Paradeep port Synergy between already purchased crude Crude matching with demand Domestic market discounts Products sale at domestic demand location price Excess product for export based on economics Variable operating cost ( Fuel & Loss) Emission norms consideration
26

Supply Chain Objectives


Maximize Corporate Profits

Profit Refinery + Profit Mktg. = Profit Corporate?


Optimize the following

Raw material Operating cost of refinery Products Logistics Cost ensuring minimal under recovery Inventory cost synchronized & optimized business process operation
Supply chain Visibility with their interdependency Quick response / Corrective actions to address internal / external contingencies

27

SCM: Integrated Approach

Crude evaluation & Procurement ?

What crude to Feed ?

What & Where To Make ?

Corporate Corporate IP IP

Demand Forecast ?

Distribution Planning ?
28

Integrated Planning Model


Supply
PORTS

Distribution R1 R2 R3
Basic Transportatio n Striucture Depots

Rn
29

Integrated Planning (IP) Model


Input - Crude Availability at Ports - Location level Demand - Desired Inventory build up / depletion - Committed Exports, Imports - Exchangeswith OMCs - Planned Shutdown schedule - Changes in product specs. - Crude Prices / Purchase Cost - Product Prices Output - Refinery wise T'put & Crude Alocation - Crude requirement for future period - Refinery wise Product Pattern - Detailed Distribution Plan Product wise, mode wise - Purchases, Exchanges - Gross Margin

Objective : Profit Maximisation


Provision of Multi-period planning
30

Model Sizes
Models Refinery Planning, RPMS Distribution Planning, SAND Integrated Planning Constraints (No of Rows) 2500 - 5500 2500 - 5500 23000 Variables (No of Columns) 6000 - 14500 8500 - 19000 63000

31

Benefits
Optimizes the whole supply chain giving higher margins and increased profitability Crude selection and allocation which takes into account product demands, refinery capabilities and effect of crudes already procured Optimal refinery production planning considering crude assay, unit capacities, product specs and demand pattern Optimal distribution planning considering transportation costs, taxes and duties and transportation constraints

32

THANK YOU

33

Overview of Industrial Practices in Planning & Scheduling -Dayanand Deshpande, 12 Jun 2009

Honeywell.com

Agenda
Supply Chain Management & Structure of Advanced Planning Systems (APS) Planning
Demand Planning & Forecasting Advanced Planning and Optimisation Distribution Planning

Scheduling
Production Scheduling Distribution Scheduling (Rail/ Road/ Ship and Pipeline Scheduling)

Implementation of APS A case study Conclusions & Outlook

2 Document control number Honeywell Proprietary

Honeywell.com

SCM- Another Short-lived Management philosophy???


The task of integrating organisational units along a supply chain and coordinating material, information and financial flows in order to fulfil (ultimate) customer demands with the aim of improving competitiveness of a supply chain as a whole
Benefit Area Reduction in Inventories
Adjust safety stock

Sample Value
Specialty manufacturer reduced inventory $15M in first year Reduction in working inventory Reduction in premium freight Customer service increase from 85% to 95% across 5000+ railcar fleet 40% reduction in forecast error by manufacturer with over 15,000 skus Production efficiencies generated additional revenues of $4M / year Estimated shift in cycle lengths of 2-8 days / line

Improved Forecasting
Statistics can reduce error 10% Collaboration can reduce error 30%

Lower Supply Chain Costs


Reduction in premium freight costs

Improved Customer Satisfaction Improved Efficiency / Capacity


Reduce process variability Improved quality Increased throughput

Non Linear Control implementation for polymers manufacturer saved millions of dollars per year Capacity increase of 8% Transition time reduction of 30% Quality variability reduction of 50% Loop scout contributed to refrigeration production increase valued at $1.7M / year at Honeywell Multi Products plant in Geismar, LA

Asset Utilization / Reliability


3-8% Improvement

3 Document control number Honeywell Proprietary

Honeywell.com

Supply Chain Management - Issues


Variability in Feedstock availability, grades and prices End-customer demand impacts RM selection and vice-versa
Refinery

Substitutions/ variety Logistics may account for Large SKUs 50% of the product cost Variability of product demand and prices Correlation with other product demand

Naphtha import/ export decisions

Naphtha

Ethylene Plant

Difficulty in matching production runs/ sequencing to PP Plant Propylene varying demand Real time Ethylene production/ yield data important

Difficult to analyse end-toend impact

PE Plant

Rail Cars

Refinery Crude Contracts & Spot market Feedstock


4 Document control number

Huge complexity, cyclic business and vulnerable to uncertainties

Primary

Secondary

Honeywell Proprietary

Honeywell.com

Decisions Supply Chain Planning & Scheduling


Time Horizon
Long term Planning
(Plant/ warehouse locn, LRP, modes of trnsprt, strategic alliances) 5- 10 years

Long Term Plans set objectives, strategies and policies

Mid term Planning


(Exchange Arrangement/ Marketing Campaigns/ Positioning Analysis )

Mon/yrs

Short term Planning


(Annual Plan/ Monthly Plan)

Weeks/ Mon

Scheduling
(Rail-Road schedule/ pipeline

Days/weeks

Details

Near Term Constraints and conditions require course corrections

schedule/production Schedule)

5 Document control number Honeywell Proprietary

Honeywell.com

Structure of Advanced Planning Systems


Procure Long-Term Produce Distribute Sell

Strategic Planning Master Planning Purch asing and MRP Production Planning
Production Scheduling

Mid -Term

Demand Planning

Distribution Planning Tx Planning & Scheduling Demand Fulfillme nt & ATP

Short-Term

(source: Supply chain Management and Advanced Planning Hartmut Stadtler & Kilger)
6 Document control number Honeywell Proprietary

Honeywell.com

Demand Planning
Starting point of supply chain planning Three stages of Demand Planning
Statistical Forecasting Judgemental (user) inputs Collaboration

Enablers Advanced Forecasting Techniques


Moving Average SMP(3) Exponential Smoothing 9 methods including Holt and Winters Box-Jenkins (ARIMA) Dynamic Regression Discrete Distributions Poison & Negative Binomial
Procure Produce Distribute Sell Long-Term

Strategic Planning

Mid -Term

Master Planning Purc hasin g and MRP Production Planning Production Scheduling Distribution Planning Tx Planning & Scheduling

Demand Planning

Short-Term

Demand Fulfillmen t & ATP

7 Document control number Honeywell Proprietary

Honeywell.com

Demand Forecasting Process


1. Gather sales history 4. Pass onto Statistical Forecasting

2. Clean data
Fix product/name changes Fix non-optimal shipment history

3. Prioritize customers
Identify ABC criteria Consolidate Tier C customers

8 Document control number Honeywell Proprietary

Honeywell.com

Forecasting Views
View all data from all sources plan, forecast, sales reps, actual, financial.

Scroll through any number of years.

9 Document control number Honeywell Proprietary

Honeywell.com

Collaborative Forecasting

Statistics as a start Entered by Sales Rep Sales Mgr auto accepted HQ revised Consensus determined

10 Document control number Honeywell Proprietary

Honeywell.com

Master Planning and Optimisation Why Now?


Shareholders demand higher profits Customers/Competitors force service improvements M&As yield more complex supply chains
More plants, distribution locations, and products

eBusiness requires accurate, automated collaboration Long-Term Enablers:


Better data (ERP, SCP) Improved supply chain models Improved LP solvers ILOG/ xpress Faster CPUs
11 Document control number

Procure

Produce

Distribute

Sell

Strategic Planning

Mid -Term

Master Planning Purc hasin g and MRP Production Planning Production Scheduling Distribution Planning Tx Planning & Scheduling

Demand Planning

Short-Term

Demand Fulfillmen t & ATP

Honeywell Proprietary

Honeywell.com

Supply Chain Optimiser


Determines most profitable customer and product mix Considers customer demands and margins, production capacity, and production/distribution costs
INPUTS OUTPUTS Customer Demand Production Capability Prices and Costs Production & Distribution Costs
12 Document control number Honeywell Proprietary

Optimized Sales Plan Optimized Production Plan

Honeywell.com

Typical Uses of SCO


Optimize & balance the supply-demands Evaluate customer trade-offs Examine equipment capacity changes Identify profit opportunities

13 Document control number Honeywell Proprietary

Honeywell.com

Distribution Planning - Highlights


Use of LP and MIP for optimising the profitability (Max sales dist. Cost) Multiplant, multiwarehouse problem construct a distribution network directly on a map Manipulating the Map Interface
Procure Produce Distribute Sell

Long-Term

Strategic Planning

Mid -Term

Master Planning Purc hasin g and MRP Production Planning Production Scheduling Distribution Planning Tx Planning & Scheduling

Demand Planning

14 Document control number

Short-Term

Demand Fulfillmen t & ATP

Honeywell Proprietary

Honeywell.com

Distribution Planning
Multi-Period Capabilities what if analysis Solution Reporting Capabilities Customized Reports

15 Document control number Honeywell Proprietary

Honeywell.com

Distribution Replenishment Planning


Predicts material needs at plants, supply points and vendor managed inventory locations Generates replenishment schedules to meet unsatisfied material demands Adjusts replenishment plans as supplies and demands vary Considers lead times, lot sizes, production schedules, policies & constraints

Demo
16 Document control number Honeywell Proprietary

Honeywell.com

Production Scheduling
Wide range of processes
Continuous, semi-continuous, batch MTO MTS mixed mode

Supports decision support & what-if analysis


Tied to and financial model

Includes embedded optimization algorithms Open architecture for unique optimization plug-ins
Procure Produce Distribute Sell

Long-Term

Strategic Planning

Mid -Term

Master Planning Purc hasi ng and MR P Production Planning Production Scheduling Distribution Planning Tx Planning & Scheduling

Demand Planning

Short-Term

Demand Fulfillme nt & ATP

17 Document control number Honeywell Proprietary

Honeywell.com

Distribution Planning and Scheduling


Distribution Planning

Road/Rai;Scheduling

Pipeline Scheduling

Ship Scheduling

18 Document control number Honeywell Proprietary

Honeywell.com

Rail Road Scheduling


Provides a single screen for Inventory Visibility

19 Document control number Honeywell Proprietary

Honeywell.com

Highlights
Complex Technical Integration (Includes interfaces from diverse sources)
ERP for Shipped Sales and Open Orders Inventory In-transit and Stock Transfers Inputs from Demand Planning Solution Inputs from Monthly Distribution Plan (integration with SAND) Inputs from customized GUIs

20 Document control number Honeywell Proprietary

10

Honeywell.com

Fleet Scheduling and Management System


Pipeline Processing POL Sea transport POD Pipeline Grid

System for Fleet Scheduling & Optimization


Fleet management LNG and other refined products Contract management Update Schedule and Annual Delivery Program (ADP) Replenishment Operational Decision Support and Re-planning of Operations

21 Document control number Honeywell Proprietary

Honeywell.com

The Fleet Scheduling Problem


Assign cargoes to ships Decide optimal visiting sequences for each ship A complex combinatorial problem
3 vessels and 5 cargoes 243 alternatives 10 vessels and 20 cargoes 100,000,000,000,000,000,000

Constraints:
Capacity Time windows (multiple) Compatibility Etc...

Rescheduling often needed

1 2 3 4 5

22 Document control number Honeywell Proprietary

11

Honeywell.com

Main Components (1)


- Electronic charts - Information of ships, cargoes, ports etc - Automatic calculation of distances - Graphical user interface - Ship positions reports by satellite - Optimization tool for fleet scheduling - Calculation for manual planning - Automatic update of ETA

23 Document control number Honeywell Proprietary

Honeywell.com

Main Components (2)


Optimise

Maintain Data

Visualise

24 Document control number Honeywell Proprietary

12

Honeywell.com

APS Solutions Case Study


Supplier

SOURCE (Procurement)

MAKE (Production)

DELIVER (Shipment & Sales)


Secondary Distribution

Primary Distribution

Customer
Sales

Integrated Planning

Crude Assay Management Crude Scheduling

Refinery Planning

Primary Distribution Planning Rail/Road Scheduling

Sec Distr Plan(LPG) Sec Distr Plan(POL)

d an em D

ng ni an Pl

Refinery Scheduling

Pipeline Scheduling

Vehicle Scheduling

KPI Management

Pr Ma ic rk e A et & na ly si s

Auto Order Generation

Implemented

Refinery MES Plant Automation

Terminal/Depot MES Terminal Automation

Back Office System Retail Automation

In progress Yet to be implemented

25 Document control number Honeywell Proprietary

Honeywell.com

Conclusions & Outlook


APS Solutions deliver tangible/ intangible benefits Challenges remain in terms of
Integration among different modules Change Management

Technologies such as RFID, SOA would be helpful in APS going forward

26 Document control number Honeywell Proprietary

13

Mathematical Programming for Scheduling of Process Operations


Dr. Munawar Abdul Shaik
B.E. (Hons.) Chemical, BITS Pilani, 1997 M.E. (Chemical), BITS Pilani, 2000 Ph.D. (Chemical Engg.) IIT Bombay, 2005 Post-Doctoral Fellow, Chemical Engg., Princeton University, 2005-2007

Assistant Professor Department of Chemical Engineering

Basic Definitions & Applications


Scheduling: Efficient management of resources over a set of units to perform a group of activities in an optimal fashion Planning: What? Scheduling: How?
Applications: Operations Research Community:
Flowshops & Jobshops Scheduling of Personnel, Vehicles etc.

Where?

How much?

Chemical Engineering:
Batch plants: Food, Pharmaceuticals, Paper products & Specialty Chemicals Continuous/Semi-continuous: Petrochemical & Refinery operations
2

People & Places


Chemical engineers have made significant contributions in Operations Scheduling in the past two decades Shah, Pantelides et al Grossmann et al Floudas et al Karimi et al Reklaitis, Pekny et al Pinto et al Ierapetritou et al Imperial College, London CMU Princeton Univ. NUS Purdue RPI Rutgers Univ.
3

Supply Chain of an Enterprise


Other Resources
Integration Challenges:
Different time scales Consistent Decisions Infrequent Revision Large-size Problems

Planning Forecasting Market Scheduling Operations Management Plant


RTO MPC

Demand/Supply

Control & Instrumentation

Vertical Integration in an Enterprise is Desirable


4

Planning and Scheduling


Multi-site Production Planning:
period 1 period 2 period T Time Horizon is longer

site 1

site 2

site N

Medium-term & Short-term Scheduling:


period t Relatively Short Time Horizons Typical Time Horizons:
Planning Medium-term Short-term 3 months to 1 year week or month hour or day

site n
5

Problem Statement
Given:
Set of products along with their demands and due dates Set of manufacturing locations Process flow sheet at each plant Equipment and storage capacities Batch processing time for each product in all stages Transition times (sequence dependent) Production, transportation, inventory or earliness, and tardiness costs

For Planning determine:


Order allocation across plants Amounts of products to be produced Length of short-term horizon

For Scheduling determine:


Optimal sequencing at each plant Start and finish times of different tasks on each unit Optimal inventory levels Optimal dispatch schedule
6

Classification of Scheduling Problems


Batch Plants Continuous/Semi-continuous Mixed production lines Serial units Parallel lines Hybrid flowshops Medium-term Scheduling Short-term scheduling Cyclic scheduling Robust scheduling Reactive scheduling

Discrete time Continuous-time formulation

Slot-based Global-event based Unit-specific-event based Precedence based


7

Classification of Scheduling Problems


Multi-product plants (job shops) Multi-purpose plants (flow shops) Without Resources With Resources (Utilities) Dedicated Storage Flexible Storage
8

Max Profit Min Make Span Min Tardiness/Earliness

Unlimited Storage No Storage / Zero-wait Finite-Intermediate Storage

Classification of Scheduling Problems


fA fB fC 1 2 3 4 A B C

Multi-product plants (flow shop) fA fB C A B fC

Multi-purpose plants (job shop)


9

Process Representation
State-Task Network (STN) Resource-Task Network (RTN) Recipe diagrams
S1 Feed A Heating i=1 S4 Hot A

STN Representation:
S8 Product 1 0.4 IntAB 0.4 Reaction 2 0.6 S5 i=4, 5 0.6 S6 IntBC

Impure E S7

0.1 Separation i=8 0.9 S9 Product 2

2 products, 3 feeds, 4 intermediates 1 heater, 2 reactors, 1 still 9 states, 8 tasks


S2 Feed B

Reaction 1 i=2, 3 0.5 S3 Feed C

Reaction 3 i=6, 7 0.2

Gantt Chart Schedule:

10

Different Time Representations


Discrete Time Representation
U1 U2 U1 U2

Continuous Time Representation I Slot based


Both start and end times of tasks have to be at an event

UN 0

UN

5 slots or 6 events
1 2 H-1 H 0 1 2 3 4 5 H-1 H

Time intervals of equal length common to all units

Time intervals of unequal and unknown length common to all units

Continuous Time Representation II


U1 U2

Global event based


Only the start times of tasks have to be at an event

U1 U2

Continuous Time Representation III Unit Specific n event based n n+1


n

UN

4 events
0 12 3 4 H-1 H

UN 0 12

Only 2 events
4 H-1 H

Events common to all units

Events different for each unit

11

Scheduling Characteristics
Profit maximization Make-span minimization Mean-flow time minimization Average tardiness minimization
UIS (Unlimited Intermediate Storage)

Performance criteria

Transfer policies

NIS (No Intermediate Storage) FIS (Finite Intermediate Storage) ZW (Zero-Wait Policy) MIS (Mixed Intermediate Storage)
12

Mathematical Model
Max Profit or Min Makespan s.t. Allocation constraints Material balance constraints Capacity constraints Storage constraints Duration constraints Sequence constraints Demand constraints Due date constraint Time horizon constraints

Mixed-Integer Linear/Nonlinear Optimization Problem

13

Solution of the Scheduling Model


Broadly Two approaches for solution: Deterministic Methods Stochastic Methods Commercial Software: Modeling Languages
GAMS ILOG OPL Studio MOSEL from XPRESSMP AMPL, LINGO etc.

Solvers
LP/MILP MINLP NLP DAEs CPLEX SBB, DICOPT, BARON SNOPT, MINOS, CONOPT MINOPT

14

Short-Term Scheduling: Batch Plants


Shaik & Floudas (2008)

RTN representation has not been explored in literature for unit-specific event-based models

Unit-Specific Event-based Continuous-time formulation for Short-Term Scheduling of Batch Plants without Resources (such as utilities) The work is extension of STN model of Ierapetritou & Floudas (1998) Improved Sequencing Constraints (for handling sequence-dependent changeovers) Alternate approach for handling dedicated finite-intermediate storage without the need for considering storage as a separate task Additional tightening constraint Limitation: Does not allow tasks to take place over multiple events

Short-Term Scheduling Model


Nomenclature Sets
I Ir R RJ RS RFIS N H Pr Dr tasks tasks related to resource r resources equipment resources material resources material resources with finite dedicated storage event points within the time horizon scheduling horizon price of resource r demand for resource r sequence independent clean up time sequence-dependent clean up time required between tasks i and i lower bound on the availability of resource r upper bound on the availability of resource r c p proportion of equipment resource produced, consumed in task i, ri 0, ri 0 p c proportion of material resource produced, consumed in task i, ri 0, ri 0

Parameters

Ermax p c ri , ri p c ri , ri

r ii ' Ermin

Short-Term Scheduling Model


Nomenclature Binary variables
w(i,n) b(i,n) E0(r) E(r,n) Ts(i,n) Assign the beginning of task i at event n Amount of material processed by task i in event n initial amount of resource r available or required from external sources excess amount of resource r available at event n time at which task i starts at event n

Positive variables

Capacity Constraints
w(i, n) Bimin b(i, n) w(i, n) Bimax
Ermin E ( r , n) Ermax
i I , n N

(1) (2)

r R , n N

Short-Term Scheduling Model


Excess Resource Balances
The amount of a resource r produced or consumed by task i is represented as:

ri w(i, n) + ri b(i, n)

p p c c E ( r , n ) = E ( r , n 1) + ( ri w(i, n 1) + ri b(i, n 1) ) + ( ri w(i, n) + ri b(i, n) ) iI r

r R , n N , n > 1
c ri c ri

iI r

(3a) (3b)

E (r , n) = E0 (r ) + ( w(i, n) + b(i, n) )
iI r

r R , n N , n = 1

The excess resource balances are more generic compared to their counterpart (material balances) in STN based models

Analysis for Equipment Resources: (keeps track of the status of a unit) J p c E (r , n) = E (r , n 1) + ri w(i, n 1) + ri w(i, n) r R , n N , n > 1
c E (r , n) = E0 (r ) + ri w(i, n) iI r
iI r iI r

r R J , n N , n = 1

A separate task is assumed for each task suitable in multiple equipment resources Implicitly represents the allocation constraint (No need to write separately)

Analysis for Material Resources: (Reduces to the material balances in STN)

Short-Term Scheduling Model


Sequencing Constraints
(i) Same task in the same unit
T s (i, n + 1) T s (i, n) + i w(i, n) + i b(i, n) i I , n N , n < N

(4)

(ii) Different tasks in the same unit:


(a) No changeovers or cleanup times:

T s (i, n + 1) T s (i ', n) + i ' w(i ', n) + i 'b(i ', n)


(b) Sequence-independent cleanup times:

r R J , i I r , i ' I r , n N , n < N

(5a)

T (i, n + 1) T s (i ', n) + i ' w(i ', n) + i 'b(i ', n) + r w(i ', n) r R J , i I r , i ' I r , i i ', n N , n < N (5b)
s

(c) Sequence-dependent changeovers:


T s (i, n) T s (i ', n ') + i ' w(i ', n ') + i 'b(i ', n ') + i'i w(i, n) H (1 w(i ', n ')) H
i "I r n '< n "< n

w(i ", n ")

r R J , i I r , i ' I r , i i ', n N , n ' N , n > n '

(5c)

Short-Term Scheduling Model


Sequencing Constraints
(iii) Different tasks in different units:
T s (i, n + 1) T s (i ', n) + i ' w(i ', n) + i 'b(i ', n) H (1 w(i ', n))
p c r R S , i ' I r , i I r , i i ', ri ' > 0, ri < 0, n N , n < N

(6)

Time Bounding Constraints


T s (i , n ) H i I , n N
i I

(7a) (7b)

T s ( i , N ) + i w( i , N ) + i b ( i , N ) H

Tightening Constraint
nN iI r

( w(i, n) + b(i, n)) H


i i

r R J

(8)

The tightening constraint provides a better LP relaxation

Objective Function
Maximization of Profit
Max Profit =

rR S

P E (r , N ) + (
r

iI r

p ri

p w(i, N ) + ri b(i, N ) )

(9)

Minimization of MakeSpan (MS)

Demand Constraints
p p E (r , N ) + ( ri w(i, N ) + ri b(i, N ) ) Dr iI r

r R S

(10)

Time Bounding Constraints


T (i, N ) + i w(i, N ) + i b(i, N ) MS
s

i I

(11)

Modified Tightening Constraint


nN iI r

( w(i, n) + b(i, n)) MS


i i

r R J

(12)

This is the model for Unlimited Intermediate storage (UIS)

Benchmark Examples
Example 1
Sundaramoorthy & Karimi (2005), and Shaik, Janak, Floudas (2006)
J1 J3 Task 1 i=1 Task 1 i=2 J2 Task 3 i=4 Task 3 i=5 J5 J4

S1

S2

Task 2

S3

S4

i=3

Problem involves 5 units, 3 processing tasks, and 4 states (1 feed, 2 int, 1 product) Variable batch sizes and processing times Finite intermediate storage (FIS) for intermediates S2 and S3 Consider two objective functions: Maximization of Profit for 3 cases of different time horizons: Case 1a: H=8 hr Case 1b: H=12 hr Case 1c: H=16 hr Minimization of Makespan for 2 cases of different demands: Case 1a: D4 =2000 mu Case 1b: D4 =4000 mu

Benchmark Examples
Example 2
Heater J1 Feed A Heating S1 Hot A S4 0.4 0.6 Reactor1 J2 Feed B S2 0.5 0.5 Reaction 1 i=2 Reaction 1 i=3 IntBC S6 Reactor2 J3 Reaction 2 i=4 Reaction 2 i=5 S8 Product 1 Product 2 0.4 0.6 IntAB S5 S9 0.9 0.1 Separation i=8 Separator J4

i=1

0.8

Reaction 3 i=6 Reaction 3 i=7

S7 Impure E

0.2

Feed C

S3

Problem involves 4 units, 8 processing tasks, 9 states (3 feed, 4 int, 2 product) Variable batch sizes and processing times Finite intermediate storage (FIS) for intermediates S4,S5,S6 and S7 Consider two objective functions: Maximization of Profit for 3 cases of different time horizons: Case 2a: H=8 hr Case 2b: H=12 hr Minimization of Makespan for the following demands: D8 =200 mu D9 =200 mu

Benchmark Examples
Example 3
Feed 3 Heater J1 Int 1 S1 Feed 1 Reaction 1 i=3 Reaction 1 i=4 Heating 1 i=1 S3 0.5 0.5 S2 Feed 2 S4 Int 2 Reactor2 J3 Reaction 2 i=5 Reaction 2 i=6 Reactor1 J2 S8 Reaction 3 i=7 Reaction 3 i=8 S10 S6 Int 4 Int 7 0.4 Mixer1 J5 Mixing i=10 Mixing i=11 Mixer2 J6 0.75 Heating 2 i=2 0.25 S9 Int 6 Product 2 S13 0.4

0.6

S5 Int 3

0.4 Int 5 Separation 0.5 i=9 S7 0.4 0.1 Separator 0.2 J4 Feed 4 S11

S12 Product 1

Problem involves 6 units, 11 processing tasks, 13 states (4 feed, 7 int, 2 product) Variable batch sizes and processing times Finite intermediate storage (FIS) for all intermediates S3 S7, S9 and S10 Consider two objective functions: Maximization of Profit for 2 cases of different time horizons: Case 3a: H=8 hr Case 3b: H=12 hr Minimization of Makespan for 2 cases of different demands: Case 3a: D12 =100 mu, D13 =200 mu Case 3b: D12 = D13 =250 mu

Benchmark Examples
Data of coefficients of processing times of tasks, limits on batch sizes of units
Task i Unit j
Unit1 Unit2 Unit3 Unit4 Unit5 Heater Reactor1 Reactor2 Reactor1 Reactor2 Reactor1 Reactor2 Separator Heater Heater Reactor1 Reactor2 Reactor1 Reactor2 Reactor1 Reactor2 Separator Mixer1 Mixer2

ij
1.333 1.333 1.000 0.667 0.667 0.667 1.334 1.334 1.334 1.334 0.667 0.667 1.3342 0.667 1.000 1.333 1.333 0.667 0.667 1.333 1.333 2.000 1.333 1.333

ij
0.01333 0.01333 0.00500 0.00445 0.00445 0.00667 0.02664 0.01665 0.02664 0.01665 0.01332 0.008325 0.00666 0.00667 0.01000 0.01333 0.00889 0.00667 0.00445 0.01330 0.00889 0.00667 0.00667 0.00667

min Bij

max Bij

(mu)
--------------------------------------------20 20

(mu)
100 150 200 150 150 100 50 80 50 80 50 80 200 100 100 100 150 100 150 100 150 300 200 200

Example 1

Example 2

(i=1) (i=2) Task2 (i=3) Task3 (i=4) (i=5) Heating (i=1) Reaction1 (i=2) (i=3) Reaction2 (i=4) (i=5) Reaction3 (i=6) (i=7) Separation (i=8) Heating1 (i=1) Heating2 (i=2) Reaction1 (i=3) (i=4) Reaction2 (i=5) (i=6) Reaction3 (i=7) (i=8) Separation (i=9) Mixing (i=10) (i=11)

Task1

Example 3

Sundaramoorthy & Karimi (2005), and Shaik, Janak, Floudas (2006)

Benchmark Examples
Data of storage capacities, initial stock levels and prices of various resources

Example 1
Storage Initial capacity stock (mu) (mu) Price ($/mu)

Example 2
Storage capacity (mu) Initial stock (mu) Price ($/mu)

Example 3
Storage capacity (mu) Initial stock (mu) Price ($/mu)

Resource

S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13

UL 200 250 UL ----------

AA 0 0 0 ----------

0 0 0 5 ----------

UL UL UL 100 200 150 200 UL UL -----

AA AA AA 0 0 0 0 0 0 -----

0 0 0 0 0 0 0 10 10 -----

UL UL 100 100 300 150 150 UL 150 150 UL UL UL

AA AA 0 0 0 50 50 AA 0 0 AA 0 0

0 0 0 0 0 0 0 0 0 0 0 5 5

UL Unlimited storage capacity AA Available as and when required

Sundaramoorthy & Karimi (2005), and Shaik, Janak, Floudas (2006)

Other models used in Comparative Study


Comparison based on our own implementation & same software and hardware STN: Ierapetritou, M. G.; Floudas, C. A. Effective continuous-time formulation for short-term scheduling: 1. Multipurpose batch processes. Ind. Eng. Chem. Res. 1998, 37, 4341. UIS Lin, X.; Floudas, C. A. Design, synthesis and scheduling of multipurpose batch plants via an effective continuous-time formulation. Comput. Chem. Eng. 2001, 25, 665. FIS RTN: Castro, P. M.; Barbosa-Povoa, A. P.; Matos, H. A.; Novais, A. Q. Simple continuous-time formulation for short-term scheduling of batch and continuous processes. Ind. Eng. Chem. Res. 2004, 43, 105. Shaik, M. A.; Floudas, C. A. Unit-specific event-based continuous-time approach for short-term scheduling of batch plants using RTN framework. Comput. Chem. Eng. 2008, 32, 260. Recipe Diagrams: Sundaramoorthy, A.; Karimi, I. A. A simpler better slot-based continuous-time formulation for short-term scheduling in multipurpose batch plants. Chem. Eng. Sci. 2005, 60, 2679.

Abbreviation used I&F

L&F

CBMN

S&F

S&K

Computational Results (UIS)


Example 1
Model Events CPU time (s)
0.05 0.01 0.02 0.01 0.01

Maximization of Profit

Nodes

RMILP ($)
2000.0 2000.0 2000.0 2000.0 2000.0

MILP Binary Continuous Constraints Nonzeros ($) variables variables


1840.2 1840.2 1840.2 1840.2 1840.2 40 20 35 10 10 215 70 85 48 68 192 86 116 69 84 84 642 274 414 176 239

Example 1a (H=8) S&K CBMN(t=1) (t=2) I&F S&F

5 5 5 4 4

13 0 7 1 1

Example 1b (H=12) S&K 9 CBMN(t=1) 9 (t=2) 9 I&F 6 7 S&F 6 7 Example 1c (H=16) S&K 12 13 CBMN(t=2) 12 (t=3) 12 (t=3) 13 I&F 9 10 S&F 9 10
a

26.83 0.23 10.32 0.03 0.19 0.02 0.23 5328.22 >67000b 1086.08 3911.14 40466.83 1.76 20.60 1.46 21.76

27176 606 21874 24 589 28 720 3408476 36297619 1642027 4087336 44252075 6596 89748 5487 91080

4481.0 4419.9 5237.6 4000.0 4857.6 4000.0 4701.8 6312.6 6381.9 7737.6 7737.6 8237.6 6601.5 6601.5 6600.9 6601.7

3463.6 3301.6a 3463.6 3463.6 3463.6 3463.6 3463.6 5038.1 5038.1 5000.0 a 5038.1 5038.1 5038.1 5038.1 5038.1 5038.1

80 40 75 20 25 20 25 110 120 105 150 165 35 40 35 40

415 130 165 76 90 106 125 565 615 225 270 295 118 132 163 182

408 162 232 115 138 130 153 570 624 319 409 448 184 207 199 199 222

1358 546 886 314 383 427 521 1895 2074 1240 1680 1848 521 590 709 803

Suboptimal solution. Relative Gap: 1.24 %b

Computational Results (UIS)


Example 2
Model Events CPU time (s)
0.07 0.01 0.03 0.28 0.03 0.23 105.5 1.82 81.95 207.43 2.16 43.73 1.79 36.28 561.58 10889.61 >67000b 331.72 4366.09 >67000c 6.19 105.64 5.29 85.67

Maximization of Profit

Nodes

RMILP ($)
1730.9 1730.9 1812.1 2305.3 1730.9 2123.3 2690.6 2690.6 3136.3 3136.3 3078.4 3551.8 2730.7 2780.2 3265.2 3315.8 3343.4 3730.5 4070.0 4409.5 3788.3 4297.9 3301.0 3350.5

MILP Binary Continuous Constraints Nonzeros ($) variables variables


1498.6 1498.6 1498.6 1498.6 1498.6 1498.6 1962.7 1860.7a 1959a 1962.7 1943.2a 1943.2a 1943.2a 1943.2a 2646.8 2646.8 2646.8a 2646.8 2646.8 2646.8a 2658.5 2658.5 2658.5 2658.5 48 32 18 26 18 26 84 56 104 144 34 42 34 42 96 108 120 120 136 152 42 50 42 50 235 104 90 115 106 135 433 170 218 258 140 165 164 193 499 565 631 248 278 308 165 190 193 222 249 114 165 216 173 224 456 189 261 321 267 318 275 326 525 594 663 298 335 372 318 369 326 377 859 439 485 672 564 783 1615 760 1238 1635 859 1046 1002 1221 1867 2119 2371 1426 1614 1802 1046 1233 1221 1440

Example 2a (H=8) S&K 5 CBMN(t=1) 5 I&F 4 5 S&F 4 5 Example 2b (H=10) S&K 8 CBMN(t=1) 8 (t=2) 8 (t=3) 8 I&F 6 7 S&F 6 7 Example 2c (H=12) S&K 9 10 11 CBMN(t=2) 9 10 11 I&F 7 8 S&F 7 8
a

4 4 13 883 10 681 88679 6449 194968 366226 6713 101415 5180 89069 288574 3438353 17270000 593182 6018234 80602289 14962 211617 12006 167306

Suboptimal solution; Relative Gap: 1.59 %b, 2.58%c

Limitation: Does not allow tasks to occur over multiple events (motivation for the Unified Model)

Computational Results (UIS)


Example 3
Model Events CPU time (s)
184.46 6.90 0.38 25.92 0.40 23.25 372.92 107.97 18.33 0.73

Maximization of Profit
Binary Continuous Constraints Nonzeros variables variables
102 121 30 41 30 41 136 165 52 52 597 264 155 190 185 226 783 348 225 267 584 343 303 377 317 391 792 457 451 465 2061 1495 875 1139 1015 1324 2789 2031 1403 1633

Nodes

RMILP ($)
2513.8 2606.5 2100.0 2847.8 2100.0 2751 3867.3 3864.3 3465.6 3465.6

MILP ($)
1583.4 1583.4 1583.4 1583.4 1583.4 1583.4 3041.3 3041.3 3041.3 3041.3

Example 3a (H=8) S&K 7 CBMN(t=2) 7 I&F 5 6 S&F 5 6 Example 3b (H=12) S&K 9 CBMN(t=2) 9 I&F 7 S&F 7
a

145888 10361 1176 57346 1074 50566 94640 47798 15871 579

Suboptimal solution

Scheduling of Refinery Operations


Dr. Munawar Abdul Shaik

Assistant Professor Department of Chemical Engineering

A Typical Oil Refinery


Crude-oil refining into useful petroleum products: LPG, gasoline, diesel fuel, kerosene, heating oil, 3 parts: Crude-oil unloading and blending Fractionation and reaction processes Product blending and shipping

(Mendez et al., 2006)

Crude-oil scheduling problem (Lee et al., 1996)


Scheduling horizon [0,H] 4 types of resources: Crude-oil marine vessels Storage tanks Charging tanks Crude Distillation Units (CDUs) 3 types of operations: Unloading: Vessel unloading to storage tanks Transfer: Transfer from storage tanks to charging tanks Distillation: Distillation of charging tanks

(Mendez et al., 2006)

Crude-oil Scheduling problem


Given Refinery configuration Logistics constraints Initial tanks inventory and composition Vessels arrival time, inventory level and composition Distillation specifications and demands (planning decisions) Determine Required operations Timing decisions Transfer volumes Minimize Cost of distilled crude-oil mixtures

(Mendez et al., 2006)

Crude-oil Scheduling problem: Example


Common logistics constraints: Only one docking station available for vessel unloadings No simultaneous inlet and outlet operations on tanks Crude distillation units can only be charged by one tank Continuous distillation

(Mendez et al., 2006)

Crude-oil Scheduling problem: Example

(Mendez et al., 2006)

Hybrid Flowshop Facility*


Consider Integration of Planning and Scheduling for an M-stage Hybrid Flowshop Plant
tanks tanks tanks

line 1

line 2
stage 1 stage 2 stage 3 stage M-2 stage M-1 stage M

Multi-level Decomposition of the Overall Problem of Integration of Planning and Scheduling


* Munawar et al., Ind. Eng. Che. Res., 2003, 2005
7

Proposed Multi-level Structure for Lube-oil scheduling


Level-1
Planning over a multi-period horizon: Order Redistribution Detailed scheduling in each period: Meeting Production Targets Operator level inventory scheduling : ISTR (Individual Tank Assignments)

Level-2

Level-3

Embedding proactive/contingency measures


8

Multi-level Structure for Lube-oil scheduling


Abstractions Level-1
Traditional timeslot usage Assumed slopping losses abstracted inventory Modified timeslot definition for slopping (Abstraction of total available compatible tank volumes) ISTR (tank assignments)

Level-2

Level-3
Increasing model granularity

Level-2: Extensions to Large problems


Performance indices Discrete variables Continuous variables No. of equations CPU time (sec) Non linear N-Z 4P3S 48 476 606 13.2 1135 5P3S 66 717 855 19.4 1609 6P3S 98 1112 1231 68.3 2388 8P3S 154 1848 1621 372.3 3667 10P3S 252 3436 2609 648.1 6251

10

Level-3: Tank assignments


Suppose: Tanks each 50 m3 capacity

11

Level-3: ISTR algorithm


Inventory level (m3) 150 100 T2 50 T1 T4 time (hr)
12

Sub-profile generation Non-overlapping zones Reuse of tanks

D
T3

C
T2 T3

B A
T3 T3

Reactive scheduling between first two levels


t=1 t=2 900 hr t=3 800 hr

Level-1

1000 hr

Level-2

Tbr=724.4 hr

Ts Tr

Nominal schedule for first period x Receding horizon for intrusion

Ts (hr) 0 1 10 25 50 75 90 99 99.9 100 101

x (hr) 0 8.678 87.8 222.13 447.59 673.89 809.88 891.52 899.66 900 900

(m3) 0 0 0 0 0 0 0 0 0 0.795 12.98

Local disturbances are attenuated locally


13

Cyclic Scheduling Problem


tanks products 1 2 N

. .
stage 1

. .
stage 2

. ... .

. .
stage M

. .

Multistage Multiproduct Continuous Plant


(Pinto & Grossmann, 1994)

14

Mathematical Formulation
Time slot stage 1 stage 2
k=1 k=2 k=1

...
k=2

k = NK

Transition processing

. . .

...

k = NK

stage M

k=1

k=2

...

k = NK

time

Time slot representation


15

MINLP model for Cyclic Scheduling


Maximize Profit =

p
i

Im im WpMi Zijkm Cinvim Ctrij Tc Tc Tc i m i j k m


(1)

1 TppikM Cinvfi iM RpiM 1 Tc TppikM 2 i k

subject to

y y
z
j

ikm

=1 =1

m k m m m

(2a) (2b) (3a) (3b)


16

ikm

ijkm

= yikm

j i

k k

z
i

ijkm

= y jk 1m

MINLP model for Cyclic Scheduling


Tspikm U im yikm 0
T

i i i

k m k m k m k m m

(4a) (4b) (4c) (4d) (5a)

Tepikm U im yikm 0
T

Tppikm U im yikm 0
T

Tppikm = Tepikm Tspikm i

Tsp
i

i1 m


i j

ijm

z ij 1 m
ikm

Tsp
i

i ( k + 1) m

Tep
i

+ ijm z ij ( k
i j

+ 1) m

k < NK m (5c)

(5b)

Tc Tpp ikm + ijm zijkm k i i j

17

Inventory Breakpoints
Tppikm Tspikm Tepikm Tppik(m+1) Tspik(m+1) Tepik(m+1) Inventory level I1im I2im

stage m stage m+1

Tppikm Tspikm Tepikm Tppik(m+1) Tspik(m+1) Tepik(m+1)

Inventory level I1im

I2im

I0im

I3im

I0im

I3im

time

time
18

Inventory Breakpoints
I 1im = I 0 im + im Rp im min Tsp ik ( m + 1) Tsp ikm , Tpp ikm k k k I 2 im = I 1im + ( im Rp im Rp i ( m +1) )max 0, Tep ikm Tsp ik ( m + 1) k k

0 I 1im Im im

I 3im = I 2 im Rp i ( m +1) min Tpp ik ( m + 1), Tep ik ( m + 1) Tep ikm k k k

0 I 2 im Im im 0 I 3im Im im Im im U im
I

I 3im = I 0 im
k

i m i i

(6) (7a) (7b)


19

Wp Mi = iM Rp iM Tpp ikM Wp Mi DiTc

Mathematical model
Variables:

y ikm {0 ,1} 0 z ijkm 1 Tsp ikm , Tep ikm , Tpp ikm , Wp Mi , T c , Im im , I 0 im , I 1im , I 2 im , I 3 im 0
Most of the Scheduling problems in Chemical Engineering result in MILP/MINLP models with large number of binary and continuous variables.

20

3P2S Scheduling Problem


fA fB fC A B C

stage 1

stage 2

Product A B C

sale price ($/ton) 150 400 650

demand (kg/h) 50 100 250

21

3P2S Problem data


stage 1 processing rate (kg/h) 800 1200 1000 intermediate storage ($/ton) 140.6 140.6 140.6 stage 2 processing rate (kg/h) 900 600 1100 final inventory ($/ton.h) 4.06 4.06 4.06

product A B C

Transition times (sequence dependent) stage 1 product A B C A 3 8 B 10 3 C 3 6 A 3 4 stage 2 B 7 0 C 3 10 -

22

3P2S Solution
stage 1
3 0 B 10.84 B 0 15.68 3 3 A 19.72 A 3 C 91.75 3 C 94.05

stage 2

23.9

Inv (ton)

Time (hr)

23

3P2S Solution
Profit = $ 442.53 / hr Cycle time = 94.05 hr Variables = 146 (18 bin) Constraints = 162 CPU time = 12.43 sec
Product A B C demand (kg/hr) 50 100 250 production (kg/h) 50 100 758

24