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 [email protected]
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.
Top Related