Benchmarking Energy Performance : From Best Practice to Next Practice

Rangan BanerjeeForbes Marshall Chair Professor

Department of Energy Science and EngineeringIIT Bombay

Invited Talk at HPCL Mumbai , 29th January 2013

Motivation

Increase in energy prices – fuels, electricity

Control costs – energy significant and growing component of production cost

Climate change – Greenhouse gas –carbon footprint

Source: www.oilnergy.com last accessed on Dec 3, 2012

Crude Oil Price Variation

Average monthly price data from July 1988 to October 2012

$/B

BL

Perform, Achieve , Trade Scheme

Source: BEE India

DEFINE AUDIT OBJECTIVES

QUESTIONNAIRE

REVIEW PAST RECORDS

WALK THROUGH / PLANT FAMILIARISATION

DATA REQUIREMENTS

MEASUREMENTS / TESTS

COMPUTE MASS / ENERGY BALANCES

ENUMERATE ENERGY CONSERVATION OPPORTUNITIES

EVALUATE ECOs

PRIORITISE RECOMMENDATIONS

DATA ANALYSIS

INSTALL MEASURES

Industry Flows

Source: Marechal, GEA

Energy Conservation Options

Waste Heat Recovery

Equipment Retrofits

Replace equipment

Cogeneration/ Efficient utility systems

Pinch Analysis/ Process Integration

Renewables

Process Improvements/ New Processes

What is Energy Benchmarking?

Comparision and target setting of energy performance – Specific Energy Consumption

Variations due to scale, age, process, raw materials, environment

Statistical Benchmarking

Zero Based Benchmarking (Thermodynamics)

Model Based Benchmarking

Thermodynamic Limits

Coal and coal products (21.5)

Crude, NGL, petroleum prod.

(13.6)

Natural gas (18.1)

Renewables (7.5)

Product

(44.6)

Loss

and

waste

(43.0)

Global

industrial

sector

Electricity (22.3)

Heat (4.6)

Total (87.6) Total

(87.6)

Coal and coal products (21.5)

Crude, NGL, petroleum prod.

(13.6)

Natural gas (18.1)

Renewables (7.5)

Product

(25.1)

Loss

and

waste

(59.2)

Global

industrial

sector

Electricity (22.3)

Heat (1.3)

Total (84.3) Total

(84.3)

Units in ExaJoules

Efficiency 51% Efficiency 30%

Energy Exergy

Source: Rosen, GEA

Non-metallic minerals

10%

Paper, pulp and print

6%

Food and tobacco

5%

Non-ferrous metals

3%

Machinery

4%

Textile and leather

2%

Mining and quarrying

2%

Construction

1%

Transport equipment

1%

Wood and wood products

1%

Iron and Steel

20%

Others

16%

Chemical and Petrochemical

29%

Industrial Energy Use Trend

Share of industrial final energy use by different sectors in 2005

World India

45%

0%

1%

1%

1%

1%

6%

7%

18%

20% Iron and Steel

Chemical and Petrochemical

Non-metallic minerals

Food and Tobacco

Paper, pulp an print

Textile and LeatherMining and Quarrying

Non-ferrous metals

Machinery

Others

Source: ETP, 2008

Data Analysis

Time Series Data

Cross Sectional Data

Pooled Time Series cum Cross sectional data

Regression Analysis (method of least squares)

Production X

En

erg

y C

on

su

mp

tio

n E

CcC

CcC

CO

nC

cC

CC

on

su

mp

tio

n E

E

Energy Consumption vs Production

E = A+ BX

Sp

ecific

En

erg

y C

onsu

mp

tio

n (

E/X

)

Production X

Specific Energy Consumption vs Production

SEC=E/X = A/X+ B

Linear first order model

ibxay

n

i

i

1

2SSE

n

i

ii yy1

2)(

2

1

))((

n

i

i ibxay

,0

a

SSE,0

b

SSE

CUSUM Technique

Cumulative Sum – Difference between baseline (expected/standard) consumption and actual consumption over a period of time

Provides trend line, Savings/ Losses

Should oscillate around zero

Helps detect impact of ECO, deterioration of plant performance

CUSUM Plot - Example

-60

-50

-40

-30

-20

-10

0

10

20

0 5 10 15 20

Series1

Source: UNIDO, 2010

Benchmarking Energy Curve for Steam crackers 2005

Source: UNIDO, 2010

Source: UNIDO, 2010

Source: UNIDO, 2010

Source: Beck , 2001

Source: UNIDO, 2010

Source: UNIDO, 2010

Glass furnace

Classification of furnace Type of firing (Cross fired /

end fired)

Raw material Batch material (like

silica, soda ash etc.)

Cullet (recycled glass)

Heat source Flame direct contact with

glass

Minimum energy requirement Heating of raw material up to

reaction temperature

Endothermic heat of reaction for batch material

Doghouse (raw material feeding section)

Throat (processed glass outlet)

Melting end

Regenerator

Checker work

Working end

Modeling practices for glass furnace

Continuum Process model Commonly used Glass furnace process in

continuum equation

Three dimensional Navier-Stokes Equation and Hottel’s zone method for radiation

Process models used mainly troubleshooting and screening variables

Limitations of process models

Data intensive inputs

Needs specialized skills and computational facilities to use

Energy performance not studied

Not easy to link operating parameters and impact on energy performance

Approach for study

Overall energy and mass balance

Study of operating glass furnaces

Identifying key operating variables

Analyzing time series data of key operating variables present in existing instrumentation

Conducting measurements for operating parameters not captured in existing

instrumentation

Literature search for furnace modelling

Refining assumptions and empirical relationships with experimental measurements

Developing mathematical furnace models with simplified assumptions for sub-processes

Solving these models for operating variables

Coupling models for understanding overall performance

Establishing relationship between dependent and independent variables empirically and analytically

Comparing measured parameters and model result

Conducting parametric of variables using validated models

Identifying areas for energy performance improvement and optimal operating strategy

Control volume

Combustion Space

Molten glass

Fuel

Batch Glass

Regenerator

Exhaust Gas Combustion Air

Control Volume 1

Control Volume 2

Control Volume 3

, , , , ,,100

, , ,

( )obh bh g l wall g g g g rk bh f bh f w w latw C

g sensi g rk bh f w

m h Q Q m h m h m h m h h

Q Q Q Q

, , , , , , , , , , , ,

, , ,,

fu comb air nonreg noncomb air nonreg air air comb reg air comb reg l wall comb g tot f f

fu air reg l reg fair nonreg

m CV m m h m h Q Q m h

Q Q QQ

, , , , , , , , , , , , , , , , , , ,f tot in f in air leak reg air leak f tot out f out l wall reg air comb reg air comb reg out air comb reg inm h m h m h Q m h h

Eq. 1

Eq. 2

Eq. 3

Mass balance of furnace

Input streams

Batch material

Cullet (recycled glass)

Raw material

Moisture

Fuel

Combustion air (from regenerator)

Air leakage (Any air other than inlet from regenerator)

Output streams

Molten glass

Cullet

Glass from raw material

Flue gas to regenerator

Combustion products

Glass reaction products

Water vapors

Air (Not reacted in combustion)

Flue gas leakage from furnace

Mass balance estimation

Estimation of flue gas formation

Based on stoichiometric Calculation of combustion

Products of combustion

Air leakage

No methodology for estimation in literature

Moisture in batch

Based on % in batch

Products of glass reaction

Based on stoichiometric Calculation of glass

Species in furnace flue gas

CO2

H2O

SO2

O2

N2

Oxygen % in flue gas (v/v dry basis)

Used as indicator for excess air control

Air leakage estimation

Furnace operates positive pressure Air leakage in local

negative pressure area

Air leakage due to higher pressure on air side

Air supplied for atomization and flame length control

Air for fuel atomization / flame control during firing and tip cooling air during non firing

Air induced by jet effect of burner

Combustion air from regenerator

Air leakage from furnace joints

Air leakage from flux line cooling

Glass melt

Energy balance for furnace

Input streams Energy from fuel

Energy from preheated combustion air

Energy from batch material

Energy from air leakage

Output streams Energy carried in glass

Heat of reaction

Sensible heat of glass

Energy carried in flue gas Energy for air leakage

Energy for batch gases

Energy for moisture

Energy for combustion air

Energy loss from walls Surface heat loss from

walls

Radiation losses (due to opening)

Energy balance glass melt

Heat of reaction for glass

Heat carried by glass

Heat carried by batch gas

Heat carried away by glass

Heat carried by batch gases and moisture

Endothermic heat of reaction for glass formation

Furnace wall losses

Glass flow direction

Flux line

Molten Glass

Zones along furnace sidewall depth

Zones along furnace melter sidewall length

Zones along furnace crown and superstructure side wall length

Furnace model input parameters

Design parameter

Design capacity of furnace

Melting area

Length to width ratio

Height of combustion volume

Refractory and insulation details

Operating parameters

Furnace draw

Type of fuel

Batch to cullet ratio

Moisture in batch

Furnace pressure

Oxygen at furnace outlet

Atomization pressure

Reversal time

Flux-line and burner tip cooling air pressure

Model flow diagram

Mass of air

Flue gas leakage

Oxygen % at regenerator outlet

Desig

n va

riable

s

Guess for total heat added

Fuel stoichiometric calculation

Glass reaction calculation

Furnace air / flue gas leakage calculations

Gap in flux line Gap near burner

Furnace operating pressure

Cooling air velocity

Number of burner

Burner air nozzle diameter

Furnace design capacity

Melting area

Furnace design details

Color of glass

Furnace geometry

Air leakage

Regenerator calculation

Flue gas outlet temperature

Heat loss from flue gas

Heat loss from regenerator wall

Oxygen % at furnace outlet

Combustion zone stoichiometric calculation

Furnace wall lossesFurnace operating

characteristics

Heat of reaction and heat carried by glass

Mass of flue gas

Heat loss from furnace area wall

Gas from glass reaction

Raw material composition

Furnace geometry calculation

Furnace design characteristics

Heat carried with glass

Heat of reaction for glass

Heat loss batch gas

Heat loss from batch moisture

Total heat added in furnace

Fuel calculationFuel calorific value

Fuel composition

Glass composition

Moisture in batch and cullet

Cullet %

Glass draw

Fuel consumptionCombustion species

Heat loss from flue gas leakage

Heat loss from air leakage

Ambient conditions

Glass outlet temperature

Port neck

Checkers packing

Glass level

1

2

5

Manual damper for airflow selection and control

6

7

Diverter damper

3

4

8

Measurement locations

Combustion air

Furnace measurement

Measurementlocation

Type of measurement

1Oxygen % , Pyrometer checkers surface temperature

2Oxygen %, Flue gas temperature

3Oxygen %, Flue gas temperature

4Oxygen %, Skin temperature

5Pyrometer checkers surface temperature

6Velocity of air at the suction of blower

7Outside wall temperature for crown and side wall

8Pyrometer glass surface temperature

Model results: Actual SEC

2.8%(118)

0.7%(30)69

1%(45)

9.7%(414)

38.2 % (1628)

2%(84)6.1%

(261)5% (212)

4.6% (198)

29.4%(1256)

33.8%(1485)

69% (2939)

Heat carried in glass

Furnace wall losses

Heat lost in moistureHeat of glass

reactionBatch gas losses

Heat loss from furnace opening

Heat lost steel superstructure

Regenerator wall losses

Heat loss from flue gas

Heat lost in cold air ingress

Heat recovery in air heating

100%(4267)

Energy introduced in furnace

From fuel 134% (5752)

Heat carried in regenerator from flue gas

Model results: Target SEC

1.7 % (63)

1.2% (45)

10.5% (390

42.7 % (1628)

1.6 % (60)7 %

(262)5.3 % (196)

5.6 % (211)

23.5 % (876)

40.5% (1510)

69.6 % (2597)

Heat carried in glass

Furnace wall losses

Heat lost in moistureHeat of glass

reactionBatch gas losses

Heat loss from furnace opening

Heat lost steel superstructure

Regenerator wall losses

Heat loss from flue gas

Heat recovery in air heating

100 % (3730)

Energy introduced in furnace

140 % (5240)

Heat carried in regenerator from flue gas

Conclusions

Target SEC estimated for 16 industrial furnaces

Effect of furnace draw on target SEC is demonstrated

0

2000

4000

6000

8000

10000

12000

14000

16000

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Furnace number

SE

C (

kJ

/kg

)

Target SEC Actual SEC

0

2000

4000

6000

8000

10000

12000

0 20 40 60 80 100 120 140 160 180 200 220 240 260 280

Draw (TPD)

Targ

et

SE

C (

kJ/k

g)

Generalized approach for model based benchmarking

Survey of existing models of process Developing

experimentation protocols

Study of actual process operation (process audit) Operating procedure

and practices Control strategy and

instrumentation Process constraints Logbook parameters

Understanding basics Defining system

boundary Writing fundamental

equations governing process

Decide assumptions Identifying empirical correlations for process

Model development Divide process into sub-

models Identify input / output

parameters for sub-models Identification of design and

operating variables Developing linkage between

process parameters and energy consumption

Experimentation

Validation of model

Refinement of model

Data from industrial process

Usage of model

Target energy estimation

Parametric analysis

Energy intensive process

Input Output Flow Diagram

Drilling/Blasting

Excavation

Transportation

Crushing/Finishing

Storage yard/dispatch

MINING UNIT BOUNDARY

INPUTS OUTPUTS

Unexcavated ore

Water

Energy requirements

Electricity

Diesel

Others

Engine oil

Lubricating oil

Finished ore

Gas emissionsCO, CO2, NOx

Dusts

Dewatering/pumping

Explosives

Waste/overburden

42

Shovel

(2.42%)

Energy usage Profileofan opencast coal mine; CIMFR study 2.5 million ton capacity

Dragline

(14.57%)Pumping

(17.85%)

Lighting

(3.01%)

Excavators(20.43%)

Dump trucks(32.52%)

Light vehicle(3.78%)

Coal handling(5.72%)

Total energy

395000 GJ(100%)

Coal handling

13%

Pumping41%

Draglines33%

Drills & Shovel

6%

Others including .

lighting7%

Electrical energy distribution pattern in Mine

Transportation58%

Excavation36%

Light vehicle

6%

Diesel consumption patternin mine

SFC= 0.152 GJ/ton

Source: ECOS, 2010

43

Variables affecting Specific fuel consumption (SFC) of Dump trucksOperating parameters

Pay Load

Distance between crusher & excavator

Speed of truck

Material handling rateMine environment

Wind speed

Mine gradient

Mine topography

Monsoon

Engine characteristics

Brake specific fuel consumption

MODEL

Speed of LoadedDump truck

Speed of Empty Dump truck

OPTIMIZATION

Distance

Minimum SFC

Specific fuel consumption

Control input

Pay Load

Fuel consumed in idling

Load, Unload time

Waiting time

Source: ECOS, 2010

Information Flow Diagram

Pce

qd

Vce

Vce

1

4,5

27

17

WG

mf,idle

WL

Vec

17

21

WE

18

ttravel

mf,ec

mf,ce

SFCdump truck

x

Mf,ij

26

BF,ec

tload,UL

2425

td,cycle

Pec

BF,ce

L

20

23

tec

tce

15

16

19

twait

Vec

Source: ECOS, 2010

Variation of SFC with pay load and material handling

Variation of diesel consumption and SFC with Payload

for 65t dump truck

Variation of SFC with handling due to increase in speed for

case of 65t dump truck

Source: ECOS, 2010

Variation of SFC with handling for Single and multiple dump trucks and also with distance

Effect of multiple dump trucks on overall SFC

Variation of SFC with distance for 65t dump truck

Source: ECOS, 2010

Generalized approach for model based benchmarking

Survey of existing models of process Developing

experimentation protocols

Study of actual process operation (process audit) Operating procedure

and practices Control strategy and

instrumentation Process constraints Logbook parameters

Understanding basics Defining system

boundary Writing fundamental

equations governing process

Decide assumptions Identifying empirical correlations for process

Model development Divide process into sub-

models Identify input / output

parameters for sub-models Identification of design and

operating variables Developing linkage between

process parameters and energy consumption

Experimentation

Validation of model

Refinement of model

Data from industrial process

Usage of model

Target energy estimation

Parametric analysis

Energy intensive process

Refinery Performance Comparison

Source: Sathaye et al, 2010

Benchmarking Energy Curve for Refineries

Source: UNIDO, 2010

Theoretical, Practical Minimum

USDOE,2006

Annual US Refining Target

USDOE,2006

Energetics,2007

Targets for 5 major processes

USDOE,2006

Steam usage- benchmarking

Solomon,2011

Optimal Cogeneration Strategy

Decisions Grid Electricity Bought/Sold

Equipment Mass Flow rates

Electric/Steam Drive

Constraints Equipment Characteristics – Min/Max

Process Steam & Electricity Loads

Grid Interconnection

Objective Function Minimise annual operating cost (Maximise

revenue)

Cogeneration

Process Steam, Electricity load vary with time

Optimal Strategy depends on grid interconnection(parallel- only buying, buying/selling) and electricity,fuel prices

For given equipment configuration, optimal operating strategy can be determined

GT/ST/Diesel Engine – Part load characteristics –Non Linear

Illustrative example for petrochemical plant-shows variation in flat/TOU optimal.

LP Steam 5. 5 b, 180 oC

Gas turbine -1

Boiler

ST

PRDS-1

PRDS-3

Condenser

Deaerator

Process Load

Process Load

40 T/h

G

1

G

4

Process Load,

60 MW

BUS

Grid

7.52 MW

SHP Steam 100 bar,500o C

HP Steam 41b,400 oC

Fuel, LSHS

9.64 T/h

WHRB-1

Supp. Firing

LSHS 5.6 T/h

Stack

20 MW

Process Load,125 T/h

Process Load,150 T/h

MP Steam 20b, 300 oC

PRDS-2

Gas turbine -2

G

1

WHRB-2

Supp. Firing

LSHS 5.6 T/h

20 MW

Fuel, HSD

5.9 T/h

136 T/h

136 T/h

131.7 T/h12.5

MW

76.2 T/h60.6 T/h

117.1

T/h

40 T/h 49.5 T/h 16.2 T/h

20 T/h

40 T/h

53.4 T/h

Make up water,357 T/h

Import Power from Grid with Cogeneration for a Petrochemical Plant

11 MW

17.6

21.6

00

5

10

15

20

25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Time hours

Imp

ort p

ow

er M

W

flat tariff TOU tariff

peak

period

demand

Export power to the grid with Cogeneration for a Petrochemical Plant

0

10

20

30

40

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Time hours

Exp

ort P

ow

er M

W

flat tariff TOU tariff

9.7 MW

Peak

period

demand

60Source: IPCC Special report 2012

Solar Fuels

61

Solar Field Components

Arun Technology CLFR Technology

Parabolic TroughScheffler paraboloid dish

62

Arun at Mahanand Dairy, Latur, India

Source: Epstein et al , 2008

Source: Epstein et al , 2007

Source: Fleiter et al, 2009

Conservation Supply Curves

CSC for Electricity savings in EU, 2030

0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

0 20 40 60 80 100 120

Cumulative Electricity Savings (GWh)

Cost

of

Saved E

lectr

icit

y (

US

2005¢ /

kW

h)

12

34

56

7

8

9

101. Automation

2. Additives

3. Optimization

4. Energy Efficient Lighting

5. Energy Efficient Motor

6. Sizing

7. Variable Spped Drives

8. New Equipment

9. Equipment Modificiation Retrofits

10. Waste Heat Recovery

Conservation supply curve for electricity savings in the Indian cement industry

Source: Rane, 2009

Summing up

Refineries – significant scope for energy efficiency improvements

Scope for model based benchmarks

Low SEC – global competiveness, lower carbon footprint

Process integration, advanced catalysts, improved control, load management

R&D for new processes, energy efficiency

Conservation Supply Curves – Demand Side Management

Pilots for Renewable Energy use

Thank You rangan@iitb.ac.in

References

Energitics, 2007: Energy and Environmental Profile of the U.S. Petroleum Refining Industry, Energitics Incorporated , Columbia, November, 2007,

Beck, T.R., 2001.”Electrolytic Production of Aluminum,” Electrochemistry Encyclopedia Electrochemical Technology Corp.

(http://electrochem.cwru.edu/ed/encycl/art-a01-al-prod.htm)

Sathaye et al, 2010: Strategies for Low Carbon Growth In India: Industry and Non Residential Sectors, Jayant Sathaye, Stephane de la Rue du Can, Maithili Iyer, Michael McNeil, Klaas Jan Kramer, LBNL, May, 2010.

USDOE, 2006: Energy Bandwidth for Petroleum Refining Processes , Prepared by Energetics Incorporated for the U.S. Department of Energy, October, 2006.

Ecofys, 2009: Developing Benchmarking Criteria for CO2 Emissions, Clemens Cremer, Joachim Schleich, Wolfgang Eichhammer, EcofysNetherlands, February, 2009.

Fraunhofer and Ecofys, 2009: Methodology for the free allocation of emission allowances in the EU ETS post 2012, Sector report for the refinery industry, November 2009.

Solomon, 2011: Solomon Associates’ Benchmarking An Insight into Energy Performance and Gaps, presented at CEE Refining and Petrochemical Meeting, October, 2011.

UNIDO, 2010: Global Industrial Energy Efficiency Benchmarking-An Energy Policy Tool, Working Paper, United Nations Industrial

Development Organisation (UNIDO), November, 2010.

Rangan Banerjee et al: Chapter 8 - Energy End Use: Industry. In Global Energy Assessment - Toward a Sustainable Future, Cambridge University Press, Cambridge, UK and New York, NY, USA and the International Institute for Applied Systems Analysis, Laxenburg, Austria.

Sardeshpande et al, 2007: Model based energy benchmarking for glass furnace, Sardeshpande, V., Gaitonde, U.N., Banerjee, R.; Energy Conversion and Management, (48)10, 2718-2738, June 2007.

Rane, 2009: Industrial DSM for Indian Power Sector.” In Proceedings of International Conference on Energy and Environment (EnviroEnergy2009, Taj Chandigarh, Chandigarh, India, March 19-21, 2009.

ECOS, 2010: Energy Performance of Dump Trucks in Opencast Mine.” In Proceedings of ECOS 2010, Lausanne, Switzerland, June 14-17, 2010.

www.oilnrgy.com (last accessed on Dec 3, 2012)

Beck, T.R., 2001.”Electrolytic Production of Aluminum,” Electrochemistry Encyclopedia Electrochemical Technology Corp.

(http://electrochem.cwru.edu/ed/encycl/art-a01-al-prod.htm)

Epstein M. et al, 2007: A 300 kW Solar Chemical Pilot Plant for the Carbothermic Production of Zinc, Journal of Solar Energy Engineering, Vol. 129, MAY 2007.