Modelling of Energy

Systems-Renewables and
Efficiency
Rangan Banerjee
Department of Energy Science and
Engineering
IIT Bombay

Seminar at Strathclyde University, June 2, 2009

Indian Examples

Analysis of wind integration into power system

PV- Battery storage Design space
Solar Water Heater Diffusion
Optimal response to time of use tariff
- process scheduling
- cool storage
-cogeneration
Benchmarking of glass furnace
Decision Support System for energy planning

Issues in grid integration

Conventional power planninghydro-thermal scheduling

How do we deal with renewables?
Capacity credit
New methodology based on Load
Duration Curve
Illustrated for Wind in Tamil Nadu

Tamil Nadu location

Tamil Nadu Grid Details

Source

Installed
capacity
(MW)

Annual
Energy
generated
(MU)

Annual
average
capacity
factor (%)

Coal

2970

21230

81.6

Gas

424

1945

52.4

Hydro

2187

6290

32.8

Firm central share

2825

17785

71.9

Wind (state + private)

3856

5270

18.6

Other renewables (solar PV,

biomass and Bagasse based
cogeneration)

556

1220

25.1

Independent power projects

(coal, lignite, diesel or gas
based)

1180

6360

61.5

Assistance from other regional

grids

519

2280

50.1

Total

14517

63370

49.8

TN Installed wind power and wind

energy generated

Hourly variation of
wind power

Monthly variation of
wind energy
generated

Capacity Credit Methodology

Variation of CC

Percentage penetration of
wind power (%)

Wind power installed

capacity (MW)

Capacity credit (MW)

5.5

500

130

11.1

1000

240

16.7

1500

350

22.2

2000

460

33.3

3000

675

44.4

4000

895

61.1

5500

1220

77.8

7000

1550

88.9

8000

1750

100

9000

1965

Methodology Wind (Micro

level)
Hourly wind speed data at
sensor height

Select major sites

Wind turbine characteristics

Hourly wind power at

sensor height
Hub height, power law
index at each site
Hourly wind power at hub
height

Installed capacity at each

site

Extrapolated hourly wind

power generated in the
state

Continue to LDC
Methodology

Hourly
wind
speed
data
Installed
capacity of
wind power

Select major sites

Extrapolated hourly wind

power generation
Effective load curve

Wind
turbine
characteris
tics

Hourly
load
curve

Divide load curve into 100

MW bins
Record number of hours in
each bin
Frequency distribution of load
over the year
Sum up to obtain annual load
duration curve
Evaluate for n discrete wind
capacities
Calculate effective base and
peak load savings from
different LDCs obtained

Input n and
n discrete
wind
capacities

Impacts on LDC

Wind power installed capacity

(MW)

Base load capacity saved

(MW)

Peak load capacity saved

(MW)

500

60

70

1000

100

355

1500

150

1105

2000

240

1265

3000

470

1475

4000

770

1625

5500

1150

1775

7000

1460

1975

8000

1630

2085

9000

1855

2125

LDC Methodology for future scenarios

Present LDC
GDP growth rate
Projected year
Elasticity
Projected LDC

Projected LDC with

renewables

Evaluate of discrete renewable

energy capacities

Compute base and peak

capacity saved

Projected renewables
installed capacity

TN Wind energy scenarios for 2021

Simulation for UK

Site: Valley, Hollyhead, Anglessy (Wales)

Wind power - load curve correlation in UK

July 0.88

Mar 0.48

Jan (-)0.51

Oct 0.39

Average correlation factor over the year = 0.38

Results for UK

No wind power
3200 MW wind (present installed
capacity)
6000 MW wind (2011 estimate)

Installed solar power

capacity

Installed wind power

capacity

Hourly solar energy

generated

Hourly wind energy

generated

Installed biomass
power capacity

Hourly biomass energy

generated

Hourly total renewable

energy generated
Effective load curve

Hourly load curve

Divide load curve into equal-sized bins

Record no. of hrs. in each bin

Frequency distribution of load over the year
Sum up to obtain annual LDC
Evaluate for n discrete renewable energy capacities
Compute base and peak load savings from diff.
LDCs

Input n and n
discrete capacities
of W, S and B

Wind resource model

WECS performance model

Solar resource model

Solar PV performance model

Biomass power generation

Micro level model

Macro level model

Utility generation model

Load model

Capacity expansion model

Economic scenario

Output: Capacity Savings

TN Solar Methodology (Micro-level)

Hourly average
insolation

Select major sites

Hourly solar power
generation at each site
Correction for panel
inclination
Correction for
temperature effect

Extrapolate based on
installed capacity at
each site to get hourly
solar power generation
for the state

Continue to LDC
Methodology

Solar module
characteristics
(efficiency vs
insolation)

Cos ()

Power coefficient
(W/deg. C)

Solar power monthly variation

Biomass power monthly variation

Installed capacity 450 MW (340 MW from bagasse cogen)

Wind and solar power required for 1000 MW

avg. peak saving

Wind and solar power installations to replace

1000 MW base power

Slope = 0.8666 which is almost equal to ratio of capacity factors (0.18/0.21 = 0.857)

Base capacity savings with wind + solar

Peak capacity savings with wind +

solar

Hybrid scenarios Impacts on

LDC

A
B
C

Impacts on LDC capacity savings

Scenario Wind
(MW)

Solar
(MW)

Biomass
(MW)

Base capacity Base capacity

saved (MW) saved (% of
installed RE
capacity)

Peak
capacity
saved
(MW)

Peak
capacity
saved(% of
installed
RE
capacity)

4000

1000

1000

1568

30.31

1252

27.22

5000

1000

1500

2137

28.66

1423

26.79

3000

3000

3000

2984

33.15

2276

25.29

Design approach
Case
study
of sample
systems

Analysis

INPUT

Compare
with
Existing
design
methods

Decisio
n
making

Integrated
design
method

Sample
design

Develop
generic
guidelines
for
design

OUTPU
T

Integrated design of Isolated power system

Load estimation

Sizing

Distribution network
using ViPOR

Load flow analysis

Yes

Is the current
location of
source gives
minimum loss
No

Relocate the source

End

Integrated designSummary
Name of
the plant

Connected
Load

(kW)

Plant
Capacity

Distribution loss
(%)

Plant capacity
factor (%)

Energy cost
Rs / kWh

Existing

Designed

Existing

Designed

Existing

Designed

Existing

Designed

Solar PV,
Rajmachi

1.4

5 kWp

4 kWp

4.6

0.5

8.3

11.5

32

25

Biomass
gasifier,
Dissoli

6.9

10 kW

10 kW

12.3

2.0

8.8

12

29-37

21-25

Biomass
gasifier,
Lonarwadi

10.7

20 kW

10 kW

14.6

2.7

5.6

14

43-54

16-25

Sizing of Photovoltaic-Battery Systems

Objective:
To arrive at the set of all feasible configurations (Array rating and Battery capacity)
to meet a given demand following a time-series simulation of the system

Schematic of the System

Photovoltaic array

Charge controller

Inverter

Load

Battery bank

DC bus

AC bus

Mathematical model
Energy balance
dQB
( PP (t ) D (t )) f (t )
dt

PP (t ) 0 AI T (t )

Power from the photovoltaic array

For a small time step, battery energy:

QB (t t ) QB (t ) ( PP (t ) D(t )) f (t )t
QB (t 0) QB (t T )
Q B (t ) 0

Br

max{QB (t )}
DOD

Hourly energy balance

Repeatability of battery energy

Non negativity of battery energy

Battery storage requirement

Photovoltaic-Battery System Sizing (Deterministic

Approach)
Inputs: Hourly solar insolation data, Hourly load data,
Photovoltaic system efficiency, Power conversion
efficiency
Estimation of the solar insolation incident on the array
System simulation to obtain the minimum array size
and the corresponding battery capacity
Calculation of the minimum storage capacity for
different array sizes greater than the minimum
Plot of sizing curve and the identification of the
design space

Graphical representation
Sizing curve for given solar insolation profile, load curve and system
characteristics

Photovoltaic-Battery System Sizing (Example)

For a remote location in Sagar island, West Bengal
Solar insolation profile

Demand profile

Averaged values for the month of December

For an average day

Sizing curve and Design-space for the

example

Mathematical Formulations

Chance constraint:

Probabilit yenergy
{D(t )
Dactual (t )} equation

Incorporating
conservation
of the
storage:

Qequivalent:
Deterministic
B (t t )
Probabilit y

QB (t )

Dactual (t ) P (t )
f (t )t
f (t )t

Generation of design space incorporating other

constraints

QB (t t ) QB (t )

( P (t ) Dactual (t ) P (t ) z )
f (t )t
f (t )t

PV-Battery System

Solar Water Heaters

Estimate potential for solar water
heaters in a given area
Develop generic framework
Diffusion of Renewable Energy
Technologies

Factors Affecting Diffusion

Of SWHS

Location- Insolation
Water Usage Pattern
Cost of electricity
Capital Cost
Reliability
Potential savings
Subsidies/ Financial Incentives

Micro Level Decision Model

(Parametric Analysis)
INPUT DATA
Water usage pattern
Location

TRNSYS

(Monthly average hourly

temperature and radiation data)

Characteristics of SWHS

Auxiliary heating requirement

(Monthly average hourly data)

Economic Analysis
MS EXCEL
Savings in Electricity Cost
Payback Period Analysis
Cost of electricity saved

Selection and sizing

of SWHS

TRNSYS (Transient System Simulation Program developed at SEL, University of Wisconsin)

Information Flow Diagram of

Micro simulation for SWHS
Weather
data

Hourly Global Solar Radiation

& Diffuse Solar Radiation
SOLAR RADIATION
PROCESSOR

Hourly ambient
Temperature

Hourly Solar Radiation

on Collector Surface

COLLECTOR

STORAGE
TANK

LOAD
(Hourly hot water
usage pattern)

AUXILIARY
HEATER

Auxiliary
heating
requirement

Model for Potential Estimation of Target Area

Target area
Weather data, area details

Single end use

point

Micro simulation using

TRNSYS
Hot water
Weather
usage pattern
data

Identification and Classification of different end uses by sector (i)

Residential (1)
Classification based on factors* (j)

SIMULATION
Auxiliary heating
requirement
Capacity of
Target
SWHS
Auxiliary
(Collector
area)
heating
Technical
Potential

SWHS
capacity
Base load
for heating

Constraint: roof
area availability

Electricity/ fuel savings

Price of
electricity
Investment
for SWHS

Economic
viability

Hospital (2)

Nursing
Homes (3)

Hotels
(4)

Others (5)

Sub-class (i, j)
Single end use point
Potential
No. of end
use points

Technical
Potential

Economic
Potential
Economic
Constraint

Market
Potential
Constraint: market
acceptance

Potential for end use sector (i = 1)

Potential Potential
for i = 2 for i = 3

Potential Potential
for i = 4 for i = 5

POTENTIAL OF SWHS IN TARGET AREA

Technical Potential (m2 of collector area)
Economic Potential (m2 of collector area)
Market
Potential (m2 of collector area)
Energy
Savings Potential (kWh/year)
Load
Shaving Potential (kWh/ hour for a monthly average day)
* Factors affecting the adoption/sizing of solar water heating systems

Potential Of SWHS
Technical potential Pij for sub-class j in sector i is

Pij f j f aj N i Psj
i.

where j denotes sub-class of end use points in sector

Psj is the simulation output for a single end use point
fj denotes fraction of the end uses
m is the total number of sub-classes.
faj is fraction of roof area availability
Ni is the number of end uses points in sector i

Pi Technical

Pij Potential for sector i is

j 1

where i denotes sector

Potential
of SWHS P(T) in the target area is
P(Technical
T )
Pi

Economic Potential
Economic potential of SWHS
P(E): subset of technical potential

P E ij ve Pij

ve = 0, if payback
acceptable limit
ve = 1, if payback
acceptable limit

period

>

maximum

period

<

maximum

Payback Acceptance Schedule

Fraction Meeting Economic
Criteria

1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0

10

12

Payback period (ye ars)

MARKET POTENTIAL

P M ij f pj Pij
fp,j is fraction of potential adopters meeting economic criteria.

Input Data For Potential

Estimation Of SWHS in Pune
Target Area

Pune

Area

138 sq.km

Total Number of households

5.17 lakhs

Number of households with more than three rooms

1.41 lakhs

Average number of persons in each household

Number of hospitals

394

Capacity range of hospitals

1-570 beds

Number of nursing homes

118

Capacity range of nursing homes

1-50 beds

Number of hotels

298

Capacity range of hotels

10-414 inmates

Number of households residing in single ownership houses

35250

Number of buildings (4 flats in each floor)

Cost of electricity
(Rs./kWh)

6 floors

1400

10 floors

880

11 floors

840

Residential

2.80

Commercial

4.00

Hot Water Usage Patterns (Pune)

70

Temperature = 40 o C

60

lit r e s/h

lit r e s/h

50
40
30
20
10
0

200
180
160
140
120
100
80
60
40
20
0

Temperature = 40 o C

10 12 14 16 18 20 22 24
Hou r of day

10 12 14 16 18 20 22 24
Hour of day

12

Temperature = 50 o C

Temperature = 50 o C

10

20

lit r e s/h

15
10
5

8
6
4
2
0

0
0

10 12 14 16 18 20 22 24
Hour of day
(c) Hospi tal (1 be d)

Temperature = 60 o C

25
20
15
10
5
0
0

8 10 12 14 16 18 20 22 24
Tim e of day (Hour)

(e ) Hote l - 1 gu e st

10 12 14 16 18 20 22 24
Hou r of day

(d) Nursi ng Home (1 be d)

30

it r e s/h

lit r e s/h

(b) Re si de nti al (2) [Nark he de , 2001]

(a) Re si de n ti al (1) [Gadgi l , 1987]

25

Monthly Average Ambient

Conditions in Pune
Ambient Temperature

30

25

25

20

20

15

15

10

10

0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month

Mani, A. (1980) Handbook of solar radiation data for India.

30

35

Temperature ( C)

Incident Solar Radiation

Monthly Average Ambient

Radiation (kWh/m /day)

Monthly Average Solar

35

Sample simulation output and potential

estimation for hospital with 5 beds
12000
10000

A mb. Temp.
Temp.at collector outlet
Temp. at tank outlet
Temp. at load

100

Solar R adiation
Solar Energy
A ux.heating

90
80
70

8000

60
6000

50
40

4000

30
20

2000

10
0
0

10

12

14

16

18

20

22

24

0
0

Hour of day

10

12

14

16

18

20

22

Hour of the day

(a) En e rgy fl ow/ S ol ar Radi ati on for a typi cal day

(b) Te mpe rature profi le s for a typical day

450
400

A nnual Electricity Savings =4290 kWh

350
300
250
200
150
100
50
0

M o nth of year

(c) Monthl y variati on in e le ctricity savings

24

Potential Estimation of Sectors

(Pune)
Technical Potential
Sector

Residential

Collector
area (m2)
Single houses
Multi-storeyed

Hospitals
Nursing homes
Hotels
TOTAL

Annual
Electricity
savings
(kWh)

Market Potential
Collector
area (m2)

Annual
Electricity
savings
(kWh)

106000

37200000

2100

740000

227400

165000000

41000

29700000

5500

5900000

1700

1600000

600

500000

300

280000

13800

15900000

9300

10740000

353300

224500000

54400

43100000

Load Curve Representing Energy

Requirement for Water Heating
1000

Typical day of January

Energy Consumption (MW)

900

Typical day of May

800

Total Consumption =14300 MWh/day

700
600

53%

500

Total Consumption = 13900 MWh/day

400
Total Consumption =760 MWh/day

300
200

Total Consumption = 390 MWh/day

100
0
0

10

12

Hour of day

14

Total Electricity Consumption of Pune

Electricity Consumption for water heating of Pune

16

18

20

22

24

Technical Potential of SWHS (sq.m.)

Achievable Potential Of SWHS For Different Payback

Periods (Pune)

400000
350000

Total Technical Potential = 353300 sq. m.

300000
250000
200000
150000
100000
50000
0
0.00

2.00

4.00

6.00

8.00

10.00

12.00

14.00

Payback Period (years)

Economic potential (limit=payback period of 5 years) =19700 m2 collector area

Framework for Potential Estimation of

Solar Water Heating Systems in a Country

Weather data
Identification of sectors and
classification within each
sector
End use details
for each subclass
Weather data
End use details

Country details
(Area, Average Weather
Data)
Locations where
Locations where
weather data
weather data
available
unavailable
Selection of base
city
Methodology for potential
estimation for a target
area
Potential of SWHS in base city
Identification of
variables for a different
Potentiallocation
of SWHS
Spatial
in different location
Interpolatio
n
Aggregation for all
locations
Potential of SWHS
country
Technical potential
Electricity savings

Potential of SWHS
in nearby area
where weather
data is not
available location
the
in the

Solar Water Heating Capacity (collector area in million

sq. m.)..

Diffusion of SWH
300

Actual installed (million sq. m.)

Potential 140 million sq. m.
Potential 60 million sq. m.
Potential 200 million sq. m.
Extrapolated Potential (million sq.m.)

250

Estimated Potential in

Potential = 200 m illion m

2092 = 199 million m

200
Potential = 140 m illion m 2

150

100

Potential = 60 m illion m 2

50

0
1990

2010

2030

2050
Year

2070

2090

Load curve of a typical day MSEB

(8/11/2000

Demand, MW

11000

source: WREB annual report-2001)

10260 MW
9892 MW

10000
9000
8000
7000

morning
peak

Evenin
g peak

6000
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

Sample Industrial Load Profile (Mumbai)

Time of Use Tariff

(MSEB-HT Ind., Jan

2002)
450
400
350
300
Paise/kWh

Peak

Peak
Partial
Peak

250

Offpeak

200
150
100
50
0
0

10

11

12
Hours

13

14

15

16

17

18

19

20

21

22

23

24

ILM Research Objective

Determine optimal response of

industry for a specified time varying
tariff develop a general model
applicable for different industries

Process Scheduling- Continuous/ Batch

Cool Storage
Cogeneration

Process Scheduling

Variable electricity cost normally not included

Flexibility in scheduling
Optimisation problem Min Annual operating
costs
Constraints Demand, Storage and
equipment
Models developed for continuous and batch
processes (Illustrated for flour mill and mini
steel plant)
Viable for Industry

Process Scheduling

Batch processes- batch time,

quantity, charging, discharging,
power demand variation (load cycles)
Raw material constraints, Allocation
constraints, Storage constraints,
Sequential Constraints, maintenance
downtime

Steel Plant Flow Diagram

St. steel Scrap mix or
Alloy steel scrap mix

Bloom caster

VD or VOD
station
Ladle Arc
furnace

40 T Melting Arc
furnace

Open store

Convertor (only for

St Steel)

Alloy steel
scrap mix

Reheat furnace

Billet caster
Bloom mill

Open store
30 T MeltingArc furnace
Open store
Reheat furnace

Bar mill

ooo
ooo

Open store
Reheat
furnace

Wire products
for final finish

Wire mill

Rods, Bars for

final finish

Flour Mill

Steel Plant Optimal Response to TOU tariff

60

Optimal with TOU tariff

Optimal with flat tariff

50
Load MW

40
30
20
10
0
2

10

12

14

Time hours

16

18

20

22 24

Process Scheduling
Summary
Example

Structure

Results

Saving

Flour Mill
Continuous

Linear, IP
120 variables
46
constraints

Flat- 2
shift 25%store
TOU-3
shift

1%
6.4%
75%peak
reduction

Mini Steel
Plant
Batch

Linear, IP
432 variables
630
constraints

Flat
TOU
Diff
loading

8%
10%
50% peak
reduction

Cool Storage

Cool Storage Chilled water operate compressor

during off-peak
Commercial case study (BSES MDC), Industrial case
study (German Remedies)
Part load characteristics compressor,pumps
Non- linear problem 96 variables, Quasi Newton
Method
MD reduces from 208 kVA to 129 kVA, 10% reduction
in peak co-incident demand, 6% bill saving

Cool Storage of Commercial

Complex
-under TOU tariff
with optimal cool storage
Load following (without storage)
250

208 kVA

kVA

200
150
100

129 kVA

50
0
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

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.

Willans Line

Fuel, HSD
5.9 T/h

WHRB-1

Gas turbine -1

136 T/h

Stack

20 MW

Fuel, HSD
5.9 T/h

Supp. Firing
LSHS 5.6 T/h
WHRB-2

Cogeneration
Example
Grid
7.52 MW

136 T/h

Process Load,
60 MW

Gas turbine -2

20 MW

BUS

Supp. Firing
LSHS 5.6 T/h

131.7 T/h
Boiler

SHP Steam 100 bar,500o C

117.1
T/h

60.6 T/h
Process Load,150 T/h

ST

Process Load,125 T/h

PRDS-1
PRDS-2
HP Steam 41b,400 oC

Fuel, LSHS
9.64 T/h

76.2 T/h

12.5
MW
G
4

20 T/h

MP Steam 20b, 300 oC

Process Load

40 T/h

49.5 T/h
PRDS-3

40 T/h

16.2 T/h

LP Steam 5. 5 b, 180 oC
53.4 T/h
Process Load
40 T/h

Deaerator
Feed water
426.5 T/h

Condenser
Make up water,357 T/h

Import Power from Grid with

Cogeneration for a Petrochemical
Plant
Import power MW

25

21.6

flat tariff

TOU tariff
17.6

20
15
10

peak
period
demand
11 MW

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

Export power to the grid with

Cogeneration for a Petrochemical
Plant
30

Peak
period
demand

20

9.7 MW

E
xport P
ow
erM
W

40

flat tariff

TOU tariff

10
0
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

Industrial Load Management

- Integrated approach
Process demand profile,
Cooling electric load
profile, Steam load profile

Captive/Cogeneration
power model

Grid tariff, fuel costs,

Grid conditions
Process load
model

Optimal
operating
strategy of
captive/
cogeneration
plant

Plant/
measured
input data

Optimal process load

schedule
Modified process
demand profile
Modified steam load
profile for process
related loads

Operating
cost
structure

Air conditioning
(cooling) load model

Optimal cool storage

Modified cooling electric

load profile

Glass furnace

Regenerator

Classification of furnace

Doghouse (raw material

feeding section)

Melting end

Raw material

Throat (processed
glass outlet)

Working end

Batch material (like silica,

soda ash etc.)
Cullet (recycled glass)

Heat source

Type of firing (Cross fired /

end fired)

Flame direct contact with

glass

Minimum energy
requirement

Checker work

Heating of raw material up

to reaction temperature
Endothermic heat of
reaction for batch material

Modeling practices for glass

furnace

Continuum Process
model

Commonly used
Glass furnace process in
continuum equation
Three dimensional
Navier-Stokes
Equation and
Hottels 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

Study of operating glass furnaces
Overall energy and mass balance

Establishing relationship between dependent and

independent variables empirically and analytically

Identifying key operating variables

Developing mathematical furnace models with

simplified assumptions for sub-processes

Analyzing time series data of key operating

variables present in existing instrumentation

Solving these models for operating variables

Conducting measurements for operating

parameters not captured in existing instrumentation

Literature search for furnace modelling

Comparing measured parameters and model result

Refining assumptions and empirical relationships

with experimental measurements

Coupling models for understanding overall performance

Conducting parametric of variables using validated models

Identifying areas for energy performance improvement

and optimal operating strategy

Control volume
Exhaust
Gas

Fuel

Batch

Combustion Air

Regenerator

Control
Volume 3

Combustion Space

Control
Volume 2

Molten glass

Control
Volume 1

Glass

&bh hbh Q&g Q&l ,wall , g m

& h m
& h m
& h m
& (hw,100 o C hw,lat )
m
1 g2 3g 1 g4 243rk 1bh4, f2 4bh3, f 1 w4 4 4
2 4 4 43
Q&g , sensi
Q&g ,rk
Q&bh , f
Q&w

Eq. 1

m&fu CV m&comb ,air ,nonreg m&noncomb ,air ,nonreg hair , m&air ,comb,reg hair ,comb ,reg Q&l , wall ,comb Q&g m&tot , f h f
14 2 43 1 4 4 4 4 4 44 2 4 4 4 4 4 4 43 1 4 4 4 2 4 4 4 3
14 2 43
&
&
Q fu
Qair ,reg
Q&l ,reg , f
Q&air ,nonreg

Eq. 2

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

Flue gas to regenerator

Cullet
Glass from raw material
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

CO2

Air leakage

No methodology for estimation in

literature

H2O

Products of combustion

SO2

Based on stoichiometric
Calculation of combustion

O2

Moisture in batch
Based on % in batch

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

N2
Products of glass reaction
Based on stoichiometric
Calculation of glass

Species in furnace flue gas

Used as indicator for
excess air control

Air leakage estimation

Furnace operates
positive pressure

Air leakage in local

Combustion air
negative pressure
from regenerator
area
Air leakage due to
higher pressure on
air side
for fuel
Air supplied for Air
atomization / flame
atomization and control during firing
and tip cooling air
flame length control
during non firing

Air leakage from

furnace joints

Air induced by
jet effect of
burner
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

Energy carried in flue gas

Heat of reaction
Sensible heat of glass
Energy
Energy
Energy
Energy

for
for
for
for

air leakage
batch gases
moisture
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

Zones along furnace
crown and superstructure
side wall length

Flux line
Zones along
furnace
sidewall depth

Zones along
furnace melter
sidewall length

Glass flow
direction

Molten Glass

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

Fuel calculation
Gap in flux line
Furnace operating pressure
Cooling air velocity
Number of burner
Burner air nozzle
diameter
Fuel consumption

Design
variables

Ambient conditions
Fuel composition
Glass composition
Moisture in batch and
cullet

Gap near burner

Furnace air / flue

gas leakage
calculations

Fuel
stoichiometric
calculation
Glass reaction
calculation

Cullet %
Glass draw

Guess for total

heat added
Oxygen % at furnace
outlet

Air leakage

Flue gas leakage

Combustion species

Combustion
zone
stoichiometric
calculation

Gas from glass

reaction

Flue gas outlet

temperature
Oxygen % at
regenerator outlet

Furnace design
characteristics
Furnace design capacity
Melting area
Furnace design details

Raw material composition

Furnace
geometry
calculation

Glass outlet
temperature
Furnace geometry

Color of glass
Furnace operating
characteristics

Fuel calorific
value

Heat loss from air

leakage
Heat loss from flue
gas leakage
Mass of air
Mass of flue gas

Regenerator
calculation

Heat loss from flue

gas

Heat loss from

regenerator wall
Heat of reaction for glass

Heat of
reaction and
heat carried
by glass

Furnace
wall losses

Heat loss batch gas

Heat carried with glass
Heat loss from batch
moisture
Heat loss from
furnace area wall

Total
heat
added
in
furnace

Furnace measurement
Measurement
location

Type of
measurement

Oxygen % ,
Pyrometer checkers
surface
temperature

Oxygen %, Flue gas

temperature

Oxygen %, Flue gas

temperature

Oxygen %, Skin
temperature

Pyrometer checkers
surface
temperature

Port neck
7

4
5
1
8

Glass level

Velocity of air at
the suction of
blower
Outside wall
temperature for
crown and side wall
Pyrometer glass

Checkers Manual damper for

packing
airflow selection and
control

2
3

Diverter
damper

Combustion air
Measurement locations

Model results: Actual SEC

Heat lost in cold air
ingress
0.7%
(30)
69
Energy
introduced
in furnace
100%
(4267)

From
fuel
134%
(5752)

Heat lost steel Heat loss from

superstructure furnace opening
1%
(45)

9.7%
(414)

2.8%
(118)

5%
(212)

Furnace wall
losses

6.1%
(261)

2%
(84)

Heat lost in
moisture
Heat of glass
reaction

Heat recovery
in air heating
Batch gas
losses

33.8%
(1485)

Heat carried in regenerator

from flue gas

29.4%
(1256)

69%
(2939)

Heat loss from flue gas

4.6% (198)
Regenerator wall losses

38.2 %
(1628)

Heat carried
in glass

Model results: Target SEC

Heat lost steel Heat loss from
superstructure furnace opening
1.2%
(45)
Energy
introduced
in furnace
100 %
(3730)

140 %
(5240)

10.5%
(390

1.7 %
(63)

5.3 %
(196)

Furnace wall
losses

7%
(262)

1.6 %
(60)

Heat lost in
moisture
Heat of glass
reaction

Heat recovery
in air heating
Batch gas
losses

40.5%
(1510)

Heat carried in regenerator

from flue gas

23.5 %
(876)

69.6 %
(2597)

Heat loss from flue gas

5.6 % (211)
Regenerator wall losses

42.7 %
(1628)

Heat carried
in glass

Conclusions

Target SEC estimated for

16 industrial furnaces
Effect of furnace draw on
target SEC is demonstrated

Generalized approach for model based

benchmarking
Survey of existing
models of process

Energy
intensive
process

Understanding basics
Defining system
boundary
Writing fundamental
equations governing
process
Decide assumptions
Identifying empirical
correlations for
process

Developing
experimentation
protocols

Model development

Divide process into submodels

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

Target energy
estimation
Study of actual process
operation (process audit)
Operating procedure
and practices
Control strategy and
instrumentation
Process constraints
Logbook parameters

Parametric analysis

Usage of
model
Data from industrial
process

Energy Planning Research

Objective
Energy Planning - Aggregated at National
Level
Sectoral Planning - Oil, Electricity
Limited Efforts in Local Level Energy
Planning
Energy critical input for development
Need for micro level energy planning
Objective: Develop a tool for improving
local level energy decisions

DSS Modules
Energy Models in Literature aggregate ,
homogenous unit, optimisation, no linkage
with decision structure
Study of decision structure, proposed an
accounting framework
Disaggregation by sector, end-uses
DST-UNDP project Bankura district in West
Bengal
Analysis District level with block as an unit
Block level- with village as an unit

Energy decisions
for district
Ratifying departmental
decisions
Zilla Parishad
Sabadhipati
Sanctioning/selecting schemes
officials
District Magistrate
Manager, DIC
DE,WBSEB/Addnl CE(Rural
Elec)

Sanctioning industries, biogas plant,

Priority list of industries.
Mouza electrification

Controller food,civil supplies

Distribution of kerosene, coal.

Lead Bank Officer

Preparation of District credit plan

District forest Officer

(DFO, working plan)
District Planning Officer

Managing forest area

Prepares district plan

Typical Energy Decisions

District Fund Allocation to blocks, Mouza

electrification, Industrial devpt., Coal elect., fuel
/ ration shops Sanctions.

Block Fund Allocation to GPs, Kerosene

allocation, industry promotion, marketing support.

Gram Panchayat Agriculture / irrigation

schemes, Co-op industry, request for fuel/ration
shop, electricity.

Household Fuel choice, Device choice.

DSS Frame Work

SECONDARY
DATA

REMOTE
SENSED
DATA

D A T A

PRIMARY
DATA

DIGITISED
MAPS

B A S E

DEVELOPMENT
PROFILE

Identify
indicators/variables
affecting energy

Trends in indicators

ENERGY DSS

SUPPLY
MODULE

DEMAND
MODULE

FUTURE ENERGY
DEMANDS BY
SECTOR /END USE

FEASIBLE ENERGY
SUPPLY SCENARIOS

IMPACT ASSESSMENT /
EVALUATION

Ground
Truthing

Remote
Sensed
images

Crop
residue
factors

CROP
RESIDUE
MODULE

Total Crop
residues
available

GIS
ANALYSIS
Estimation of
crop areas
Digitized
Maps
Map Showing
Landuse
Classification

Secondary
Data

Estimation of
wood volume
correlations

Areal
extent of
forest, nonforest

Sub
Classification of
Sal, and Mixed
forests
ACCOUNTING
MODULE

Average fuelwood
density non-forest
(Survey / literature)

Fuelwood
non-forest

FUELWOOD
MODULE

Total
Fuelwood
available

Fuelwood
avail from
forest areas
Height, Density
measurements
Standing wood
estimates forest
areas

ACCOUNTING
MODULE

% of
Fuelwood

Drivers for Demand

Scenarios
Drivers

Residential
cooking

Residential
Noncooking

Agriculture

Industrial

Population
Low-MedHigh
(L-M-H)

Population
(no of
Households)
Low-MedHigh
(L-M-H)
Income NT-T
Electrification
NE-ME-AE

Rainfall
Low-Med-High
LR MR HR
Irrigation
NI LI- MI HI
Crop Pattern
NTC/TC1/TC2
Pump Electric
NE/ME/AE

Growth
Low-Med-High
LG/MG/HG
*Rice mills linked
to agricultural
scenarios

18

108

3+

Income
No
Transition(NT)
/ Transition(T)
No of
combinations

Select Scenario cases

Populatio
n

Income

Village
Electrificati
on

SC1

NT

NE

SC2

SC3

SC4

SC5

Rain

Irrigatio
n Land

Crop
Patter
n

Pump
Electrifie
d

Ind.
Growth

HR

NI

NTC

NE

LG

AE

LR

HI

TC2

AE

HG

ME

MR

MI

TC2

ME

MG

AE

HR

HI

TC1
TC2

AE

HG

ME

HR

MI

TC1

ME

MG

Electricity Demand in
2005
557

600

GWh in 2005

500

421

400

328

283

300
200

154
136 GWh

100
0
SC1

SC2

SC3

SC4

SC5

Demographics

Appliance ownership data

1.

Number of Households (HH)

2.

% of HH in each income class

Technology
Characteristics
1.
2.

Total no of
Appliances in Area

Rating (kW)
Efficiency
Electricity consumption in

Usage pattern
1.
2.

by each income class

Residential sector. (Including

Daily variation
Seasonal Variation

the load curve) for each

season/Annual

Number of
Electric
Pumpsets
Usage and
seasonal
variation

Agriculture Load

Technical
specifications
Number of
Commercial
shops

Technical
Characteristics
and Usage

Appliance
ownership

Total
Appliances

Commercial
Load

Total Electricity
demand.
Daily load
Curve
Summer/Winter

Village Electrification example

High I ncome
13%

Middle
I ncome
23%

Below
Poverty
33%
Above
Poverty
31%

Income class distribution in Rajamele in

1994.

70

Load (kW)

60
50
40
30
20
10
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour of the day.
Residential Electricity Demand
Industry demand

Agriculture Demand
Total Load

Daily load curve for Rajamele village in Summer in 2005

DSS Capability/Use

Energy Quantification- from data already

available/collected by Govt
Assess impacts of development paths
Quantify future commercial fuel demands
elec/coal/LPG/kerosene
Rural Electrification- Load Profile
Estimation,Sizing, Option Selection, Impact
Assessment
Spatial Representation of results
Classification of areas as biomass surplus/deficit

Summing Up

Examples illustrate variety of

optimisation/ simulation models for
energy sector
Decision context and model
formulation critical
Reality check - Applicability
Generalisation important but..
Data intensity and uncertainty

Acknowledgment

S. Ashok Indu Pillai

Vishal S.
Ph.D - 2003Ph.D - 2008 Ph.D. - 2008

U.N.
Gaitonde
Faculty

Sandeep Phulluke, Sagar Kasar,

Abhijit Bhure,Amit Khare, Jagdish
DSS Project Staf
(2004-2006)
Balkrishna Surve
Project Assistant

Santanu B.
Arun P.
Ph.D. - 2009 Faculty

Mel George A.
Manojkumar M.V.
M.Tech - Ongoing M.Tech - Ongoing

Thank you

References

Ashok.S and R. Banerjee , An Optimisation model for Industrial Load management, IEEE Trans on Power
Systems, Vol.16, No. 4, Nov.2001, pp 879-884.
Ashok S and R. Banerjee , Optimal Operation of Industrial Cogeneration for Load Management, IEEE Trans
on Power Systems, Vol. 18, No. 2, MAY 2003.
Ashok.S and R.Banerjee, Optimal cool storage capacity for load management , Energy, Vol. 28, pp 115126, 2003.
Vishal S, U.N.Gaitonde,R Banerjee., Model based energy benchmarking for glass furnace, Energy
Conversion and Management, Vol.48, pp 2718-2738, 2007.
R. Banerjee, A. B. Inamdar, S. Phulluke, B. Pateriya, Decision Support System for Energy Planning in a
District, Economic and Political Weekly, Vol. 34, No. 50, pp 3545-3552, 1999.
R.Banerjee, Rahul Pandey, Abhijit Bhure and Sandeep Phulluke ,Electricity Demand Estimation for Village
Electrification , Proceedings of National Renewable Energy Conference , IIT Bombay, Nov. 30-Dec.2 ,
2000.
Rahul Pandey, R.Banerjee, Abhijit Bhure and Sandeep Phulluke ,Framework for Design and Evaluation of
Electric Supply Options for a rural area, Proceedings of National Renewable Energy Conference , IIT
Bombay, Nov. 30-Dec.2 ,2000.
Arun P., Santanu Bandyopadhyay and R. Banerjee, Sizing curve for design of isolated power systems,
Energy for Sustainable Development,Volume XI, No. 4, December 2007.
Indu R. Pillai and R. Banerjee, Methodology for estimation of potential for solar water heating in a target
area, Solar Energy, Vol.8, No.2, pp 162-17, 2007.
George, R. Banerjee,Analysis of impacts of wind integration in the Tamil Nadu grid, in press, Energy Policy
UK Wind speed data: GWEFR Cyf Hourly-mean wind speed datasets for sites in the European
Wind Atlas, available at http://www.gwefr.co.uk/datasets.htm
UK load curves: Demand Data, UK National Grid, available at