WIND GENERATION OPERATIONS CHALLENGES AND … RE and ERCOT... · MAPE = MEAN ABSOLUTE %ERROR....

WIND GENERATION OPERATIONS

CHALLENGES AND EXPECTATIONS

OPERATIONS PLANNING STAFF

ERCOT

PUBLIC 2

Objectives• List two challenges in forecasting wind in Texas.

• List three actions a Wind facility experiencing icing issues should take.

• Given an example of telemetry being received from a renewable

resource, determine if it is an issue and whether the QSE representing

this resource should be contacted to rectify the issue.

PUBLIC 3

Wind Generation and Wind

Forecasting in ERCOT

PUBLIC 4

816 977 1,173 1,385 1,8542,875

4,785

8,0058,916 9,400 9,604

10,40711,065

12,470

15,76416,631 16,631 16,631

2,872

6,086 6,455248

2,291

3,952

116

816 977 1,173 1,3851,854

2,875

4,785

8,005

8,9169,400

9,60410,407

11,065

12,470

15,764

19,751

25,008

27,038

0 MW

5000 MW

10000 MW

15000 MW

20000 MW

25000 MW

30000 MW

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018

Cumulative MW Installed IA Signed-Financial Security Posted IA Signed-No Financial Security

The data presented here is based upon the latest registration data provided to ERCOT by the resource owners and can change without notice. Any capacity changes will be reflected in current and subsequent years' totals. Scheduling delays will also be reflected in the planned projects as that information is received. This chart reflects planned units in the calendar year of submission rather than installations by peak of year shown.

Financial security posted for funding interconnection facilities does not include CREZ security deposits, which are refunded to the Interconnecting Entity when an IA is signed.

ERCOT Wind Installations by Year (as of July 1, 2016)

Wind Growth Potential

PUBLIC 5

Wind Regions within ERCOT

Panhandle:

3,171 MW

South:

976 MW

West:

9,572 MW

North:

1,116 MW

Coast:

2,041 MW

PUBLIC 6

Wind Records

TIMELOAD

(MW)

WIND GENERATION

(MW)

PENETRATION AT

RECORD TIME

02/18/2016 21:20 35,528 14,023 39.47%

03/23/2016 01:10 27,245 13,154 48.28%

PUBLIC 7

Wind Generation Forecast

• AWS Truepower computes a Short Term Wind Power Forecast (STWPF) for all

Wind-powered Generation Resources (WGR) in ERCOT for a rolling 168 hour

timeframe.

• This process uses information from multiple systems, including:

– RARF (Resource Asset Registration Form) (ex. site geo-location, met tower geo-location)

– EMS (Energy Management System) (ex. Resource Status, telemetered site specific

meteorological data)

– Outage Scheduler (ex. WGR derate start/stop dates, derate values)

• When submitting COPs, Wind QSEs must use this STWPF for the applicable

168 hours timeframe. (COP HSL must be <= STWPF.)

– Note that NPRR 785, upon approval will only require Wind QSEs to update a COP only

when wind farm operating conditions dictate COP HSL to be lower STWPF.

*Note that for a WGR going through commissioning process, ERCOT will start providing STWPF to the QSE once WGR receives approval to

Generate into the ERCOT Grid (Part II of the Commissioning check list)

PUBLIC 8

WPF Process Inputs

Source Data

Registration

System

1. Resource Name

2. Location of wind farm (latitude and longitude or equivalent for the center point of wind farm)

3. Location of the meteorological tower (latitude and longitude or equivalent)

4. Type (manufacturer/model) and number of turbines

5. Turbine hub height(s) above ground level with associated number of turbines

6. Manufacturer’s power curve (capability curve)

7. Resource Commercial Operation Date

Energy

Management

System (EMS)

1. Most recent Resource (wind farm) status with date/time

2. Most recent MW output of wind farm with date/time

3. Most recent wind speed and direction at hub height from one meteorological tower with date/time

4. Temperature and barometric pressure at 2 m above ground level on the meteorological tower

Telemetry Value

(Sent every 5

minutes from

EMS)

1. MW Average

2. Wind Speed (mph)

3. Wind Direction (degrees)

4. Temperature (ºC)

5. Barometric Pressure (mbar)

6. HSL Average

7. Num of Turbines ON

8. Num of Turbines Off

9. Num of Turbines Unknown

10. Curtailment Flag

Outage

Scheduler

1. Turbine Outage/Unit Derate Value (HSL)

2. Turbine Outage/Unit Derate Start

3. Turbine Outage/Unit Derate Stop

PUBLIC 9

NPRR 785

• Synchronizes Short Term Wind Forecast (STWPF) with Current Operating Plan

(COP) of Wind Generation Resources (WGRs) and Photovoltaic Generation

Resources (PVGRs)

– ERCOT will automatically prepopulate every WGR and PVGR’s COP with most recent

forecast for the next 168 hours.

– QSEs representing WGRs or PVGRs will be required to either submit a COP with the

automatically prepopulated forecast or submit a lower number (when wind farm operating

conditions dictate COP HSL to be lower STWPF).

PUBLIC 1

0

• Relatively concentrated geographic region for wind installation (lack of spatial

diversity)

• Texas weather can change rapidly

• Severe changes in wind generation output caused by different types of extreme

weather (large wind ramps)

– Frontal system, trough, or dry line/Thunderstorms/Low-level jets/Weakening pressure

gradients/Strengthening pressure gradients

• There is an approximation of hypothetical wind speed-power curve to the real

one

Hypothetical

Power Curve Actual Wind Farm

Power

*The steep part of the power curve from 4-12 m/s is where power

increases strongly with speed.

Challenges in Forecasting Wind

PUBLIC 1

1

Wind Forecast Errors

0

1000

2000

3000

4000

5000

6000

7000

8000

0%

1%

2%

3%

4%

5%

6%

7%

8%

Jul-15 Aug-15 Sep-15 Oct-15 Nov-15 Dec-15 Jan-16 Feb-16 Mar-16 Apr-16 May-16 Jun-16 Jul-16

Mo

nth

ly M

ean

Es

tima

ted

Un

cu

rtaile

d P

ow

er O

utp

ut [M

W]

Mo

nth

ly M

ean

Ab

so

lute

Pe

rce

nt

Err

or

of

Insta

lle

d C

ap

acit

y

WIND POWER FORECAST SYSTEM-WIDE Mean Absolute Percent Error

Monthly Mean Estimated Uncurtailed Power Output [MW] Day-Ahead 1430 STWPF

Day-Ahead 1430 COP HSL Hour-Ahead STWPF

Hour-Ahead COP HSL

––

PUBLIC 1

2

Wind Forecast Errors (Day-Ahead)

––

2013

2014

2015

6.0%5.7%

6.7%6.0%

6.9%

5.2%5.6% 5.8%

5.1%

4.0%

0.0%

2.0%

4.0%

6.0%

8.0%

10.0%

12.0%

14.0%

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

MA

PE

DAY AHEAD WIND FORECAST PERFORMANCE

2013 2014 2015 2016

MAPE = MEAN ABSOLUTE %ERROR

PUBLIC 1

3

Wind Forecast Errors (Hour-Ahead)

––

2013

2014

2015

3.84%3.52%

3.81%

4.26%4.51%

3.88%3.64%

3.98%

3.40%

2.70%

0.0%

1.0%

2.0%

3.0%

4.0%

5.0%

6.0%

7.0%

8.0%

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

MA

PE

HOUR AHEAD WIND FORECAST PERFORMANCE

2013 2014 2015 2016

MAPE = MEAN ABSOLUTE %ERROR

PUBLIC 1

4

Wind Generation Cold

Weather Preparation

PUBLIC 1

5

Icing Impacts

Source: Wind Turbines in Cold Climates: Icing Impacts and Mitigation Systems, http://www.springer.com/us/book/9783319051901

PUBLIC 1

6

Wind Units Failure in Cold Winter

• Frigid Temperatures Exceeds Turbine Limits

‒ Wind turbines are typically designed to operate within ambient air temperatures of -15°C/-

20°C (5F/-4F)

‒ Wind turbines have an automatic shutdown feature to protect components if that range is

exceeded.

• Blade Icing

‒ Icing on wind turbines affects three different aspects simultaneously

• the design (aerodynamics, load, control system, and material)

• the safety (ice throw, unbalanced rotor spinning, over-power, and fatigue)

• performance (annual energy output, wind measurements, and design life duration)

‒ Other impacts

• wind sensors—rendering ineffective wind-measuring equipment

• increase noise levels and generally decrease a turbine’s cost-effectiveness.

PUBLIC 1

7

Blade Icing

PUBLIC 1

8

Safety

PUBLIC 1

9

Inclement Weather Plan

Preventative Maintenance

Schedule

Emergency Plan

Communication

Cold Weather Preparation

PUBLIC 2

0

Cold Weather Extreme Package• This ensures that wind turbine operates in temperatures as low as -30C (-22F),

and in survival mode without operation, at temperatures as low as -40C (-40F)

(one example).

• ERCOT is aware of at least one vendor that offers an extreme cold weather

package. Resource Entities should check with their manufacturer.

Minimum temperature (standard)

operational / survival

Standard weather package -15°C/-20°C (5F/-4F)

Cold weather package -30 °C/-40 °C (-22F/-40F)

PUBLIC 2

1

Cold Weather Packages

• “Cold Weather Packages” extend temperature ranges by heating components

such as the nacelle space, yaw drive and pitch motors, and the gearbox, slip

ring, controller and control cabinet, and battery.

• Heating could also prevent another insidious cause of turbine failure.

– When a turbine is not running, oil that is stationary in radiator passages can quickly cool,

and its viscosity can increase.

– Even if wind turbines are not being used, an important lesson worth learning is that the

turbines should be cycled online to provide flow of cooling oil.

– All cooling equipment for radiators on wind turbines should also be disabled for cold weather

events.

PUBLIC 2

2

Icing Prevention Technologies• Passive icing prevention methods

– Rely on the physical properties of the blade surfaces to prevent ice accumulation.

– An example is application of an anti-adhesive coating on the blade such as teflon.

– Another approach takes advantage of the heat absorbing capacity of dark colored surfaces

and consists in the use of black coated blades.

• Active de-icing methods

– Consist of thermal, chemical and impulse de-icing

– In thermal de-icing, electrical elements, similar to the one found on the rear window of a car,

can be used to warm and melt the ice accumulation off the blades.

• In a comprehensive wind turbine icing prevention approach, sensors that could

detect the build-up of ice on the rotor could be considered.

PUBLIC 2

3

Wind Operations During Icing

Conditions

PUBLIC 2

4

Challenges Created by Icing

• Icing creates a mismatch between the forecast and actual production.

• This creates an error in the forecast which then was provided to WGRs.

• This error may be used in the COP which feeds into RUC.

– Could lead to ERCOT counting on more wind than WGRs are able to generate.

PUBLIC 2

5

Icing Event Dec 26-27, 2015• Freezing temperatures on December 26, 2015 in the panhandle region of Texas

marked the beginning of icing conditions that lasted throughout the duration of

December 27.

• By 3:00 AM on December 27, the wind forecasting error had reached about

1700 MW. Icing was continually reported throughout the day and forecasting

error reached nearly 4500 MW by the hour ending HE0800.

• The outage scheduler reported new WGRs related derates and forced outages

totaling around 6000 MW by the end of the day.

PUBLIC 2

6

Wind Outages 12/26 – 12/27

PUBLIC 2

7

West Wind Forecasts Over Time – HE8

Observed: 2,285

Last COP: 5,864

6 hour: 6,937

PUBLIC 2

8

West Wind Forecasts Over Time – HE19

00

Observed: 1,352

Last COP: 1,774

6 hour: 3,110

12 hour: 5,592

PUBLIC 2

9

West Wind Hour Ahead Summary

0.0

1,000.0

2,000.0

3,000.0

4,000.0

5,000.0

6,000.0

7,000.0

8,000.0

9,000.0

10,000.0

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

Forecast Error STWPF Aggr COP RT Aggr Wind-Output

PUBLIC 3

0

Panhandle Wind Hour Ahead Summary

0.0

500.0

1,000.0

1,500.0

2,000.0

2,500.0

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

Forecast Error STWPF Aggr COP RT Aggr Wind-Output

PUBLIC 3

1

Forecast vs. COP Comparison

0.0

1,000.0

2,000.0

3,000.0

4,000.0

5,000.0

6,000.0

7,000.0

8,000.0

9,000.0

10,000.0

1 3 5 7 9 11 13 15 17 19 21 23

West Wind Hour Ahead Summary

Forecast Error

STWPF Aggr COP

RT Aggr Wind-Output

0.0

500.0

1,000.0

1,500.0

2,000.0

2,500.0

1 3 5 7 9 11 13 15 17 19 21 23

Panhandle Wind Hour Ahead Summary

Forecast Error

STWPF Aggr COP

RT Aggr Wind-Output

PUBLIC 3

2

Notice the disparity between the last forecast value before the adjustment period and the observed value in HE 8 and 12

versus in HE 19. Delayed outages caused these errors in earlier hours which were then minimized in HE 19. The faster

these outages are entered, the faster they are incorporated to the forecast.

Forecast Performance

0

500

1000

1500

2000

2500

HE 3 HE 8 HE 12 HE 19

5 Hour Forecast MW

Last Forecast Before Adjustment Period MW

Observed

Panhandle Wind West Wind

0

1000

2000

3000

4000

5000

6000

7000

8000

HE 3 HE 8 HE 12 HE 19

5 Hour Forecast MW

Last Forecast Before Adjustment Period MW

Observed

PUBLIC 3

3

Unit’s STWF with no Outage Updates

Forecast NEVER

Converged to zero

WRONG

PUBLIC 3

4

Unit’s STWF with Outage Updates

Forecast Converged

to zero

CORRECT

PUBLIC 3

5

Telemetry During Icing Conditions

Number of

online turbines

Number of

offline turbines

MW output

*Remember this is under icing

conditions.

While the NTON line does

decrease, it does not fall as the

unit’s MW goes to zero, but is

very delayed.

This is too slow.

WRONG

PUBLIC 3

6

Telemetry During Icing Conditions

Number of

online turbines

Number of

offline turbines

MW output

*Remember this is under icing

conditions.

The NTON line falls as the unit’s

MW goes to zero.

This is very good.

CORRECT

PUBLIC 3

7

Addressing the Wind Turbine Icing

• Use the following when a WGR is experiencing an icing related event,

– Call the ERCOT Control Room and notify them.

– Continuously update the telemetry for number of turbines on and off (NTON and NTOFF) as

conditions on the ground deteriorate.

– If the outage/derate is expected to last greater than 2 hours, submit this into ERCOT’s

Outage Scheduler.

• In comparison to turbine telemetry, Outage submissions have a greater impact in

correcting the forecast.

3.1.4.5 Notice of Forced Outage or Unavoidable Extension of Planned or Maintenance Outage Due

to Unforeseen Events

(2) Any Forced Outage that occurs in Real-Time must be entered into the Outage Scheduler if it is to

remain an Outage for longer than two hours.

PUBLIC 3

8

RRGRR-0003• A survey was conducted in 2015 to gather additional information to gauge the

operational limits of the various WGRs in ERCOT.

‒ Maximum Operating Temperature (Fahrenheit (°F))

‒ Minimum Operating Temperature(Fahrenheit (°F))

‒ High Wind Speed Cut-Out (meters per seconds (m/s))

‒ High Wind Speed Cut-Out time (minutes)

‒ High Wind Speed Cut-Out Reset (meters per seconds (m/s))

‒ High Wind Speed Cut-Out Reset Time (minutes)

• This information is now part of the RARF and is fed into the wind forecast

received from AWS.

PUBLIC 3

9

Current and Ongoing

Telemetry Concerns

PUBLIC 4

0

Meteorological Telemetery

Telemetry Name Data Unit

Wind Speed (MPH) MPH

Wind Direction (DEG) Degrees

Temperature (TEMP) ºC

Barometric Pressure (PRES) mbar

Num. Of Turbines ON (NTON)

Num. Of Turbines OFF (NTOF)

Num. Of Turbines Unknown (NTUN)

Integer (Count of turbines)

PUBLIC 4

1

Telemetry Examples (from 9/6/2016)

41

PUBLIC 4

2

0

20

40

60

80

100

120

140

0

5

10

15

20

25

30

35

40

06

:00

06

:17

06

:35

06

:53

07

:10

07

:28

07

:46

08

:03

08

:21

08

:39

08

:56

09

:14

09

:32

09

:49

10

:07

10

:25

10

:42

11

:00

11

:18

11

:35

11

:53

12

:11

12

:28

12

:46

13

:04

13

:21

13

:39

13

:57

14

:14

14

:32

14

:50

15

:07

15

:25

15

:43

MW HSL NTON NTOF

• Wind farms interpret the NTON flag differently.

• The NTON is meant to describe the number of turbines available for production,

and has nothing to do with the actual output.

– The left image is taking the unit’s MW value into account - Wrong

– The right image is expected method for computing and telemetering this data.

Number of Turbines Online (NTON)

0

5

10

15

20

25

30

35

06

:00

06

:16

06

:33

06

:50

07

:06

07

:23

07

:40

07

:57

08

:13

08

:30

08

:47

09

:04

09

:20

09

:37

09

:54

10

:11

10

:27

10

:44

11

:01

11

:17

11

:34

11

:51

12

:08

12

:24

12

:41

12

:58

13

:15

13

:31

13

:48

14

:05

14

:22

14

:38

14

:55

15

:12

15

:28

15

:45

MW HSL NTON NTOF

CORRECTWRONG

0

5

10

15

20

25

30

350

6:0

00

6:1

60

6:3

30

6:5

00

7:0

60

7:2

30

7:4

00

7:5

70

8:1

30

8:3

00

8:4

70

9:0

40

9:2

00

9:3

70

9:5

41

0:1

11

0:2

71

0:4

41

1:0

11

1:1

71

1:3

41

1:5

11

2:0

81

2:2

41

2:4

11

2:5

81

3:1

51

3:3

11

3:4

81

4:0

51

4:2

21

4:3

81

4:5

51

5:1

21

5:2

81

5:4

5

MW HSL NTON NTOF

WRONG

0

20

40

60

80

100

120

140

0

5

10

15

20

25

30

35

40

06

:00

06

:17

06

:35

06

:53

07

:10

07

:28

07

:46

08

:03

08

:21

08

:39

08

:56

09

:14

09

:32

09

:49

10

:07

10

:25

10

:42

11

:00

11

:18

11

:35

11

:53

12

:11

12

:28

12

:46

13

:04

13

:21

13

:39

13

:57

14

:14

14

:32

14

:50

15

:07

15

:25

15

:43

MW HSL NTON NTOF

The NTON line stays constant despite

the unit’s MW value going to zero.

CORRECT

PUBLIC 4

6

0

20

40

60

80

100

120

140

00

:00

00

:37

01

:15

01:5

302

:31

03:0

90

3:4

70

4:2

505

:03

05:4

106

:19

06

:57

07

:35

08:1

308

:51

09:2

91

0:0

6

10

:44

11:2

212

:00

12:3

81

3:1

6

13

:54

14:3

215

:10

15:4

81

6:2

6

17

:04

17:4

218

:20

18:5

81

9:3

5

20

:13

20

:51

21:2

922

:07

22

:45

23

:23

MW HSL

MW exceeding HSL telemetry• MW telemetry must never exceed the HSL.

• When a WGR is UNCURTAILED its MW and HSL values should be exactly the

same.

WRONG

PUBLIC 4

7

0

20

40

60

80

100

120

00

:00

00

:41

01

:22

02

:03

02

:44

03

:26

04

:07

04

:48

05

:29

06

:10

06

:52

07

:33

08

:14

08

:55

09

:36

10

:18

10

:59

11

:40

12

:21

13

:02

13

:44

14

:25

15

:06

15

:47

16

:28

17

:10

17

:51

18

:32

19

:13

19

:54

20

:36

21

:17

21

:58

22

:39

23

:20

MW HSL

MW exceeding HSL telemetry• MW telemetry must never exceed the HSL.

• When a WGR is UNCURTAILED its MW and HSL values should be exactly

the same.

CORRECT

PUBLIC 4

8

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

0

10

20

30

40

50

60

70

80

19:0

0

19

:03

19

:07

19:1

1

19:1

5

19

:19

19:2

2

19:2

6

19:3

0

19

:34

19:3

8

19:4

1

19

:45

19

:49

19:5

3

19:5

7

20

:00

20

:04

20:0

8

20:1

2

20

:16

20:1

9

20:2

3

20:2

7

20

:31

20:3

5

20:3

8

20:4

2

20

:46

20:5

0

20:5

4

20

:57

MW HSL Curtailment Flag

WGR Under Curtailment

• ERCOT nodal protocols and Business Practices dictate that when under

curtailment a WGR’s HSL must represent actual power generation potential.

WRONG

• HSL should never

lower dramatically

under curtailment

scenarios.

PUBLIC 4

9

WGR Under Curtailment• ERCOT nodal protocols and Business Practices dictate that when under

curtailment a WGR’s HSL must represent actual power generation potential.

CORRECT

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

0

20

40

60

80

100

120

140

160

12

:00

12

:28

12

:56

13

:24

13

:53

14

:21

14

:49

15

:17

15

:46

16

:14

16

:42

17

:10

17

:39

18

:07

18

:35

19

:04

19

:32

20

:00

20

:28

20

:57

21

:25

21

:53

22

:21

22

:50

23

:18

23

:46

00

:14

00

:43

01

:11

01

:39

02

:08

02

:36

03

:04

03

:32

MW HSL Curtailment Flag

• HSL should

never lower

dramatically

under

curtailment

scenarios.

PUBLIC 5

0

WGR Under Curtailment• WGRs must follow ERCOT Base Point when under curtailment

CORRECT

100

105

110

115

120

125

130

135

140

145

150

18:0

0

18:0

3

18:0

7

18:1

1

18:1

5

18:1

9

18:2

2

18:2

6

18:3

0

18:3

4

18:3

8

18:4

1

18:4

5

18:4

9

18:5

3

18:5

7

19:0

0

19:0

4

19:0

8

19:1

2

19:1

6

19:1

9

19:2

3

19:2

7

19:3

1

19:3

5

19:3

8

19:4

2

19

:46

19

:50

19

:54

19

:57

MW HSL Base Point

PUBLIC 5

1

Questions?