ENABLING HIGH PENETRATION OF WIND ENERGY IN THE PACIFIC NORTHWEST - - PowerPoint PPT Presentation

ENABLING HIGH PENETRATION OF WIND ENERGY IN THE PACIFIC NORTHWEST OPERATING ENVIRONMENT

Cesar Silva Monroy – Graduate Student

• Prof. Richard D. Christie

University of Washington Energy Seminar – Sept. 25, 2008

Slide 2

OVERVIEW

• 1. Introduction
• 2. Wind Integration Issues
• 3. High Penetration Examples
• 4. Measures of the Problem
• 5. Strategies for Integration
• 6. Conclusions

Slide 3

• 1. INTRODUCTION
• Wind power is growing rapidly in the U.S. and worldwide

because:

– Increased concern on climate change and as part of efforts to reduce greenhouse gases – Government policies such as production tax credit (PTC) to promote the use of renewable energy technologies – Public initiatives, also known as renewable portfolio standards (RPS), such as the Washington Energy Independence Act (I-937) – Wind power technology advances such as higher wind capacity factors and lower O&M costs

Source: 20% Wind Energy by 2030, DOE Source: Annual Report on U.S. Wind Power Installation, Cost, and Performance Trends: 2006, DOE

Slide 4

• 2. WIND INTEGRATION ISSUES
• Wind’s variability and uncertainty makes wind power different from
• ther conventional power plants

– Feb 24, 2008: Approximately 1.1 GW of industrial and commercial loads served by ERCOT (Texas) were shed as emergency response to drop in West Texas wind energy production from 1700 MW to 300 MW in about 2½ hours – June 1, 2006: Portugal suffered a 700 MW wind energy decrease, corresponding to 60% of capacity in 8 hours – Jan. 8, 2005: There was a decline of 2000 MW, 83% of total capacity, in a period of 6 hours in Denmark – Dec. 24, 2004: More than 4000 MW of wind energy decrease within 10 hours which corresponds to 58% of capacity in North Germany

Slide 5

• 2. WIND INTEGRATION ISSUES
• When compared to conventional generation, wind’s stochastic

nature results in:

– Lower controllability – Higher variability – Lower predictability

5 10 15 20 25 30 35 40 45 3/12 3/19 3/26 4/2 4/9 Time Wind speed (mph)

Data source: Idaho National Lab

Slide 6

• 2. WIND INTEGRATION ISSUES
• Power system operation before wind power:

Generation = Load

– Generation is controllable and schedulable – The load follows a pattern highly predictable – Power system paradigm: Generation changes to match the load variations

Slide 7

• 2. WIND INTEGRATION ISSUES
• Power system operation with wind power:

– Wind generation is not controllable and difficult to schedule – Wind is a paradigm shift for the power system: Controllable generation changes to match the load minus the stochastic generation variations

Controllable Generation = Load - Stochastic Generation

• Today, power systems use their imbalance capability employing

controllable generation to cope with wind power variations

Source: 20% Wind Energy by 2030, DOE

Slide 8

• 2. WIND INTEGRATION ISSUES
• What happens when the controllable generation cannot follow

these variations?

• There will be incidents such as violations to control

performance standards or system instability

• How can these incidents be avoided?
• Identifying wind penetration limits

– Economic limit : cost-effectiveness of new wind power plants is driven down by the high costs of interconnecting new wind projects, making them economically not viable – Control limit : the power system does not have sufficient resources to meet ACE criteria such as CPS1 and CPS2 – Physical limit : wind power variations could result in power system collapse $/MWh Wind (%) penetration lE lC lP Integration cost

Slide 9

• The relationship between cost and wind penetration levels is

particular to each system

• We want to allow for wind integration at lowest possible cost
• How to increase controllable resources at lowest possible cost,

shifting the wind integration cost curve and limits?

• New paradigm:

Controllable resources must match the variations from non- controllable resources

• Increasing controllable resources will allow higher penetration

levels

• 2. WIND INTEGRATION ISSUES

$/MWh Wind (%) penetration lP Integration cost lE lC lE

Slide 10

• 2. WIND INTEGRATION ISSUES
• As wind penetration increases, the way traditional generators
• perate will change

– Base load units – Marginal units – Peaking units

• How does this change in operation practices affect current

controllable generation?

• How does it affect other infrastructure in the power system?

Load MW Hours

8760

Ppk

Slide 11

• 3. HIGH PENETRATION EXAMPLES
• a. Denmark
• b. Ireland
• c. Spain
• d. Western U.S.

Slide 12

• 3. HIGH PENETRATION EXAMPLES
• The Western U.S. is a leader in the U.S. in terms of total wind

installed capacity (California: 2,439 MW, Washington: 1,163 MW and Oregon: 882 MW)

• An addition of new wind projects is expected in order to meet

new demand for renewable energy as required by RPS in WA, OR and CA

• Increase wind installed capacity across the Western

interconnected system

• How does this situation compare to other examples of high

wind penetration?

• Europe presents many scenarios with different characteristics

Source: Annual Report on U.S. Wind Power Installation, Cost, and Performance Trends: 2006, DOE

Slide 13

• 3a. DENMARK
• Wind energy was supplying over 20% of the energy demand in

Denmark by the end of 2006

• Denmark is interconnected to the Nordic countries, Germany

and France

• Nordic countries have a lot of hydro-generation (Norway: 99%,

Sweden: 50%, Finland: 17%)

• The mix generation resource in France is dominated by nuclear

energy (78.3% of total electric power generation)

• Germany also has wind generation spread along the country,

with a higher concentration in the Northeast region (5.7% wind penetration)

• Denmark can export wind energy when the wind blows, and

import it when it does not

• The regional wind penetration level is high, but the system

penetration level is low

• Wind is geographically spread over the country

Slide 14

• 3b. IRELAND
• Wind supplies nearly more than 7% of Ireland’s energy demand

according to data from 2007

• Small electric system with low level of interconnection to other

systems

• The primary fuels used to generate electricity are natural gas,

coal and peat

• Renewable energy will continue to grow thanks to energy

policies with targets such as:

– Supply of 15% and 33%of total electricity consumption using renewable sources by 2010 and 2020 respectively – 500 MW installed tidal energy by 2020

• The Irish government has also set targets for increasing the

interconnection capacity of the island

Slide 15

• 3c. Spain
• Wind power plants reached 16% of the total intalled capacity in

Spain and supplied 10% of the electric energy in 2007

• Spain’s electric grid is managed by a single entity or system
• perator (Red Electrica de España)
• Installed capacity (86 GW) is far greater than the peak demand

(45GW)

• Wind is geographically diverse
• Generation mix is composed of mainly combined cycle units

(24%), hydro (19%), nuclear (9%) and coal (13%)

REE presentation – Wind forecasting workshop, Portland, OR 2008

Slide 16

• 3d. WESTERN U.S.
• Most of wind power generation is concentrated in areas like

western Washington and Oregon

• Load centers are generally located far from the wind generation

areas

• Wind power seems to be located in the wrong side of the

transmission bottlenecks

• The generation mix is dominated by hydro-generation in the

Pacific Northwest

• California presents a different generation mix, with natural gas

being the primary fuel

• All of these conditions make the Western interconnection

system a very particular case for wind integration

Slide 17

• 4. MEASURES OF THE PROBLEM
• a. Control Performance Standards (CPS1 and CPS2)
• b. Loss of Load Probability (LOLP) and Expected

Unserved Energy (EUE)

• c. Reserves

Slide 18

• 4a. CONTROL PERFORMANCE STANDARDS
• Control Performance Standards (CPS) 1 and 2 were developed

under the old power system paradigm

• Intended to identify areas with poor control performance
• The Area Control Error (ACE) is used as the basis of analysis
• Statistical analysis that deals with the body of the error

probability distribution, and not with the tail

• Are these standards too strict in an environment with high

stochastic penetration? Should they be relaxed?

Slide 19

• 4b. LOSS OF LOAD PROBABILITY AND EXPECTED

UNSERVED ENERGY

• LOLP is a measure for comparing service reliability and is “the

fraction of time during which loss of load may be expected to

• ccur during any future period”
• EUE is defined as “expected amount of energy not supplied by

the generating system during the period of observation, due to capacity deficiency”

• These indicators are based on probability analysis
• Some important questions are:

– Are they adequate for stochastic resources? – Do they need to be redefined in the presence of new controllable resources such as demand response and wind curtailments?

Slide 20

• 4c. RESERVES
• Current methods for reserve estimation are operating fine for

present levels of wind integration

• As wind penetration level increases it is necessary to capture

the need for increasing reserves

• Are current methods adequate for this task?
• A too low value of reserves results in excessive interruptions
• A reserve value too high results in excessive costs
• Do current reserve definitions capture all the controllable

aspects of the power system?

Slide 21

• 5. STRATEGIES FOR INTEGRATION
• a. Conventional deterministic generation
• b. Specialized deterministic generation
• c. Wind curtailments
• d. Real-time customer demand response
• e. Wind prediction
• f. Energy storage technologies

Slide 22

• 5. STRATEGIES FOR INTEGRATION
• New paradigm:

Controllable resources must change to match the variations from non-controllable resources

• Controllable resources might include:

– Conventional deterministic generation – Specialized deterministic generation – Wind curtailments – Real-time customer demand response – Wind prediction – Energy storage technologies

• Each of these technologies can help meet the controllability

requirements of wind, but at what cost?

Slide 23

• 5a. DETERMINISTIC GENERATION
• Power systems are designed to handle load variations using

well established means such as automatic generation control (AGC) and economic dispatch

• Demand-generation balance capability can be used under the

assumption that a wind power plant output resembles a negative load

• Default approach in industry today because it uses existing

units

• As more wind generation projects are built, more regulating

reserve is required, keeping reserve capacity off the market and displacing energy production to less efficient units

Slide 24

• 5b. SPECIALIZED DETERMINISTIC GENERATION
• Generation units installed specifically to provide regulation for

non-controllable generation such as wind

• Characteristics:

– High ramp-rates – Medium to high capital costs – High O&M costs

• Their operation is optimized for controllability
• Examples of this type of generation are:

– Natural gas combustion engines – Gas turbines

Slide 25

• 5c. WIND CURTAILMENTS
• Capping output power and/or ramp rates of wind generation

plants would increase integration of wind power plants into systems with deterministic generation response capability

• Non-symmetrical response given that down rates of wind

cannot be controlled

• This approach would have effects on the income of wind power

plants

• The cost of ramp rate curtailments to wind power plants is lower

than output power curtailments

Slide 26

• 5d. REAL-TIME CUSTOMER DEMAND RESPONSE
• Transfer of real-time decision making capabilities to the end

users on the electric grid is now possible thanks to advances in communications and information technologies

• Pacific Northwest National Laboratories (PNNL) demonstrated

that it is possible to modulate the load demand using a price signal:

– Within a short period of time – At a small capital cost – Small effects on customer satisfaction – Using residential electric water heaters and thermostats, commercial building space conditioning and municipal water pump load – Reduced peak loads, increasing transmission capacity and lowering energy prices

• Demand response also has the potential to follow the output

power variations of new wind power plants

Slide 27

• 5e. WIND FORECASTING
• Complex methods such as neural networks are employed for

forecasting long term wind variations

• Very complex schemes combine meteorological data, power

system data and wind speed and weather forecast

• There is an inability to predict large wind speed changes in the

short term

• A potential solution to this problem is upstream sensing
• What benefits to the power system can be extracted from this

information?

• Does the value of short term wind speed prediction justify its

implementation?

Slide 28

• 5f. STORAGE
• Energy stored would be used to compensate for fluctuations in

the output power of wind to create a deterministic power output for the combined power plant

• Their cost-benefit ratio remains a concern due to high capital

costs, low round trip efficiencies and, in some cases, high self discharge rates

• Among these storage technologies are:

– Pumped hydro – Compressed air energy storage – Flow batteries – Hydrogen storage

Slide 29

• 6. CONCLUSIONS
• The ability to cope with wind power variations using existing

controllable generation is limited in any power system. How can limits be determined?

• At current wind capacity growth rates in our region, these limits

might be reached in the near future

• The Western electric system has very particular characteristics

that demand further research in order to identify the most cost- effective alternatives for wind integration

• Use of new controllable resources will help increase the wind

penetration further in order to keep up with the demand for wind energy

Slide 30

QUESTIONS?