1

Long‐term Demand Forecasting

11th Capacity Building Programme for Officers of Electricity Regulatory Commissions

Long term Demand Forecasting

&

Power Procurement Planning

for Distribution Utilities

1

for Distribution Utilities

(Uttar Pradesh) Dr. Anoop Singh

Dept. of IME, IIT Kanpur1

Need for Demand Forecasting &Power Procurement Planning (PPP)

• Long‐term Forecast – Planning for capacity addition/power procurementg p y /p p– Upgrade transmission facilities

• Medium‐term Forecast– Planning for power procurement– Design tariff structure– Demand‐side management– Time of use pricing

2

p g

• Short‐term Forecast– Merit order dispatch– Minimising deviation from schedule – Decision making for short‐term power procurement– Optimising use of renewable energy

2

2

Objectives* 

• Demand projection for the power requirement in Uttar Pradesh

h l bl d d d f• Assess the available, proposed and expected power procurement fromconventional and renewable energy sources

• Optimize the power procurement to meet future peak load & energyrequirement

• Develop a power procurement scenario with a mix of long‐term, medium‐t PPA d h t t t

3

term PPA and short‐term power procurement

* ‐ The study was conducted on behalf of UPPCL so as to enable the utilities to meet their obligations underPower for All programme.

3

Trend  • Study the past growth pattern

Methodology – Four Stages

1. Projection of electrical energy  demand1. Projection of electrical energy demand

• Study category‐wise connected load, electricity consumption and growth pattern

End Use method

• Forecast considering economic change 

Econometric Models

AnalysisStudy the past growth pattern energy demand

2. Load profile analysis & projection

3. GAMS  based optimisation 

4

•Inference from historical load profile

•Account for projected solar addition and DSM 

•Account for demand profile influenced by supply

2. Load profile analysis & projectionmodel

4. Choosing power procurement  strategy

4

3

• Develop optimisation model considering

3. GAMS  based optimisation model

Methodology (contd…)

1. Existing & candidate plants’ quantum and prices

2. Projected solar capacity addition and DSM activities

3. With and without short‐term procurement  

4. Generator & contract specific constraints

• Defining different power procurement scenarios

5

• Estimation of social cost and utility cost in different economic and power procurement scenarios

• Identify optimal power procurement strategy

4. Choosing power procurement  strategy

5

Projection of Electrical Energy Demand

6 6

4

90%

100%

Electricity Demand Pattern Across States  

1200

Per Capita Electricity ConsumptionCategory‐wise Electricity Consumption (2014‐15)

50%

10%

8%

28%

10% 11% 5%12%

7%

11%

7%

3%

3%

11%

15%5% 3% 3%

8%

6%

9%4%

16%

37% 38%

27%25%

18%

25%

23%

28%21%

3%

18% 22%

30%33%

36%

25%38%

16%18%

30%

40%

50%

60%

70%

80%

90%

672717 734

779819

884 914957

1010

346 372 387 412450 450 472 502400

600

800

1000

kWh

7

41%

27%

50%

17% 21% 19%24% 23% 20%

28%39%5%

11% 7%

0%

10%

20%

Domestic Commercial Industrial (LV & MV)Industrial (HV) Agriculture So: CEA (Gen. Review 2016)

341 346 372

0

200

India Uttar Pradesh

So: CEA (Gen. Review 2006‐077 to 2014‐15) 7

Electricity Consumption vs Economic Activity 

ChhattisgarhDelhi

GujaratHaryana

Karnataka MaharashtraRajasthan

TN

1000

1500

2000

umption 

kWh)

Bihar

MPRajasthan

UP

0

500

0 20000 40000 60000 80000 100000 120000

Per Cap

ita Consu

of Electricity (k

Per Capita SGDP  (Rs. at 2004 prices)

Elasticity ‐ 0.78

2011‐12450

500

ty 

h)

Uttar Pradesh 

8

2006‐072007‐08

2008‐092009‐10

2010‐11

2012‐13

300

350

400

450

14000 14500 15000 15500 16000 16500 17000 17500 18000 18500 19000

Per Cap

ita Electricit

Consumption (kWh

Per Capita SGDP  (Rs. at 2004 prices)

Elasticity ‐1.16

8

5

Power Sector Scenario in Uttar Pradesh

41.30 %

70 %

%30 00

35.00

40.00

45.00Category‐wise Consumption (in %) 

14000160001800020000

nds

Number of Consumers CAGR ‐ 8.55%

7.40 %

22.7

18.30 %

0.00

5.00

10.00

15.00

20.00

25.00

30.00

2010‐11 2011‐12 2012‐13 2013‐14 2014‐15 2015‐16

Domestic Commercial Industrial Agriculture

02000400060008000

100001200014000

2010‐11 2011‐12 2012‐13 2013‐14  2014‐15  2015‐16 2016‐17

Thousan

Source –Data from UPPCL

4500050000

ds

Total Connected Load (kW) CAGR ‐8.04%

70000

Uttar Pradesh Electrical Energy Consumption(in GWh) Utility Only

9

05000

1000015000200002500030000350004000045000

2010‐11 2011‐12 2012‐13 2013‐14  2014‐15  2015‐16 2016‐17

Thousand

Source Data from UPPCL

30000

35000

40000

45000

50000

55000

60000

65000

Source ‐ CEA, General review 2016

CAGR ‐ 8.16%

9

Trend Analysis

• Useful for preliminary estimate for forecast

Forecasting Methods

• Uses time factor, assuming pattern of demand in the past will continue into future

• Does not explain underlying factors– Demographics– Economic factors Consumer behaviour and price sensitivity

1 0

– Consumer behaviour and price sensitivity– Government initiative, tech. development  and RE induction

• Easy to use and understand

10

6

Forecasting Methods (Contd…)

End‐use method 

d h d l l– Focuses on various end‐uses in the residential, commercial, agriculture and industrial sectors

– Aggregate energy demand is summed over different end‐uses in a sectors

– Effective method when there is lack of adequate past data

1 1

– This method requires a high level of detail on each of the end‐uses

– Does not take in account demographics and socio‐economic factors

11

Forecasting methods (Contd…)

Econometric Analysis

• Utilises information from historical data along with factors that• Utilises information from historical data, along with factors that influence electricity demand

– Demographic factors (urbanisation, number of households & size, connected load etc.)

– Economic factors (GDP/per capita income, industrial growth etc.)

C ti b h i b t

1 2

– Consumption behaviour by consumer category 

– Price sensitivity

• Estimate econometric relationship and use it to predict future demand 

12

7

Econometric Model

1 3 13

Data & Data Sources • CEA ‐General Review (2003‐04 to 2014‐15)

i. Category‐wise Connected Load, No. of Consumers and Consumption all states

ii. Per capita electricity consumption

iii. Number of pump‐sets energized, Mid year population

• CSO, MOSPI

i. State Gross Domestic Product at constant price (base year 2011) 

• PFC Reports 

i. Weighted average price of electricity (base year 2011)

• Tariff Orders

i. Power procurement cost

1 4

p

• UPSLDC

i. Unrestricted demand, rostering, emergency rostering, electricity generation etc.  

• UPPCL 

i. U.P. number of consumers, connected load and consumption category‐wise

ii. PPA Information's and rate of electricity 

iii. CS3 & CS4 reports 14

8

Forecasting Scenarios

• High Growth Scenario

• Realistic Growth Scenario

• Medium Growth Scenario

• Low Growth Scenario

1 5 15

Projected Per Capita Electricity Consumption in U.P

10111000

1200

953

854

721

400

600

800

1000

kWh

High Growth

Realistic Growth

Medi mGro th

1 6

0

200

200

3‐04

200

4‐05

200

5‐06

200

6‐07

200

7‐08

200

8‐09

200

9‐10

201

0‐11

201

1‐12

201

2‐13

201

3‐14

201

4‐15

201

5‐16

201

6‐17

201

7‐18

201

8‐19

201

9‐20

202

0‐21

202

1‐22

202

2‐23

202

3‐24

202

4‐25

202

5‐26

202

6‐27

Medium Growth

Low Growth

16

9

Projected Electrical Energy at Bus‐bar for Utility Only

244238

300000

244238

230398206808

175223

50000

100000

150000

200000

250000

GWh

High Growth

Realistic Growth

1 7

0

50000

200

3‐04

200

4‐05

200

5‐06

200

6‐07

200

7‐08

200

8‐09

200

9‐10

201

0‐11

201

1‐12

201

2‐13

201

3‐14

201

4‐15

201

5‐16

201

6‐17

201

7‐18

201

8‐19

201

9‐20

202

0‐21

202

1‐22

202

2‐23

202

3‐24

202

4‐25

202

5‐26

202

6‐27

Medium Growth

Low Growth

17

Results Comparison With Other Reports

Note: For utilities only

Compassion Projected Energy ( 19th EPS vs Estimated Value) GWh

FYCEA Econometric model results (IIT Kanpur)

19 EPS Realistic High Medium Low* Without Captive Generation

Note: Energy sold

19 EPS Realistic High Medium Low

2016‐17 108070 114512 114512 114512 1145122021‐22 150797 163562 166115 153757 142298

2026‐27 195323 227838 244238 206808 175223

Projected Total sales (In MU)FY PFA  Econometric Model ∆ %

1 8

Note: Energy sold * Without Captive  and losses2016‐17 83789 92882 11%

2017‐18 95131 101267 6%2018‐19 103173 110511 7%2019‐20 116385 120706 4%2020‐21 126046 130958 4%2021‐22 136700 141753 4%

18

10

Load Profile Analysis & Projection

1 9 19

Low Duration Curve 

2 0 Source ‐ Night Report ,UPPCL  20

11

Uttar Pradesh – Load profile

13500

15500

17500

MW

Monthly Avg. Load ProfileJan,17

Feb,17

Mar,17

14500

15500

16500

17500 June month avg. load profile

June, 2013

June, 2014

MW

9500

11500

13500

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

M

Apr,17

May,17

Jun,17

2 1 Source ‐ Night Report ,UPPCL 

9500

10500

11500

12500

13500

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

June, 2014

June, 2015

June, 2016

June, 2017

M

21

Projected Load Profile – High Growth Scenario

Annual Load Profile (High Growth Scenario)

26206

30058

20000

25000

30000

MW

2026-27

2025-26

2024-25

2023-24

2022-23

2021-22

2020-21

2 2

10000

15000

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

2019-20

2018-19

2017-18

2016-17

2015-16

Hour22

12

Projected Load Profile – Realistic Growth Scenario

2833730000 Annual Load Profile (Realistic Growth Scenario)

2026 27

24704

16000

18000

20000

22000

24000

26000

28000

MW

2026-27

2025-26

2024-25

2023-24

2022-23

2021-22

2020-21

2019 20

2 3

10000

12000

14000

16000

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

2019-20

2018-19

2017-18

2016-17

2015-16

Hour23

Projected Load Profile – Medium Growth Scenario

28000.0 Annual Load Profile (Medium Growth Scenario)

2026 27

16000.0

18000.0

20000.0

22000.0

24000.0

26000.0

MW

2026-27

2025-26

2024-25

2023-24

2022-23

2021-22

2020-21

2 4

10000.0

12000.0

14000.0

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

2019-20

2018-19

2017-18

2016-17

2015-16Hour

24

13

Projected Load Profile – Low Growth Scenario

24000

Annual Load Profile (Low Growth Scenario)

2026 27

16000

18000

20000

22000

MW

2026-27

2025-26

2024-25

2023-24

2022-23

2021-22

2020 21

2 5

10000

12000

14000

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

2020-21

2019-20

2018-19

2017-18

2016-17

25

GAMS  Based Optimisation Model

2 6 26

14

Input Variables

Minimise social cost of power supply

2 7 27

GAMS Simulation for Different Power Procurement  Scenario 

Demand Load Profile High, Realistic, Medium &  Low Growth Scenario

With or Without Short‐term Contracts

Solar & DSM ‐Realistic Targets 

Fi d P iti Fl ti P iti

Solar & DSM –Policy Targets 

For candidate 

2 8

Fixed Position 

Fixed All PlantsFixed 3 Plants Float Panki @ (2023,24,25)

Floating Position  

Float All PlantsFloat 3 Plants  & cancel 1 (H, 

J, O, P)

Similar Scenarios as for Realistic 

Targets 

Plants only

28

15

GAMS Output

• Demand & Generation Chart

PLF f Pl t• PLF of Plants

• Optimal position for candidate plants

• Utility Cost  & Social Cost  

2 9 29

Demand & Gen. Curve for Realistic Growth Scenario (Fix All Without Considering STPP)

Load Shed

30

Existing Generators

Candidate Generators

10

15

20

25

rocured in thousand M

W

3 0

Solar

Existing Generators

0

5

1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19

2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027

Power Pr

Solar Existing Generators Candidate Generators Load Shed Demand

30

16

Demand & Gen. Curve for Medium Growth Scenario(Fix All Without Considering STPP)

Load Shed25

30

ExistingGenerators

Candidate Generators

10

15

20

25

Power Procured in thousand M

W

3 1

Solar

Existing Generators

0

5

1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19 1 7 13 19

2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027

P

Solar Existing Generators Candidate Generators Load Shed Demand

31

Choosing Power Procurement  Strategy

3 2 32

17

Conclusions & Key Observations

• Most optimal strategy ‘All Float’ case

• Certain existing plants/PPAs have low PLFg p /

• Progress on DSM and Solar capacity addition should be reviewed for medium‐term power procurement

• Planned and effective measures should be in place for metering, meter reading, billing and collection

• Periodical Power Procurement Analysis (at least every three years)

3 3

• Extension of Time of Day (ToD) tariff for all large consumers

• fair and transparent competitive bidding

• Adopt a state‐level UMPP model

33

Disruptive Changes in Future

Open Access

f l Rooftop Solar

Retail competition

Metro & Electric Traction

Electric vehicles

Smart Grid

3 4

Smart Grid

Storage 

Franchisee (with exist clause for power procurement?)

34

18

Thank You

www.iitk.ac.in/ime/anoops

3 5

anoops@iitk.ac.in

35