Wind and Solar Integration in Colorado: Challenges and Solutions Colorado Rural Energy Agency Energy Innovations Summit Energy Storage for Intermittent Resources: Is the “Silver Bullet” on the Horizon? Keith Parks, Senior Analyst October 14, 2011

Agenda Colorado Wind & Solar Energy Today Renewable Energy Integration Challenges  Solar  Wind Is Storage the Solution?  Natural Gas Price  Carbon Policy  Integration Costs

Colorado Wind and Solar Today PSCo – 1735MW SPS – 736MW NSP-MN – 1591MW Xcel Energy – 4062MW Wind = 1735MW 1159 turbines Solar = 131MW* 8652 installations *includes 27.4MW Utility Scale Solar PV at two installations 2x30MW utility scale solar facilities to be operational by end of 2011 Additional 400MW of wind energy installed by end of 2012

Solar Energy Integration PV is the dominant solar technology on the system and will be for the near-term PV is dramatically affected by cloud cover – variability can occur in seconds. Peak production from fixed- axis installations match poorly with system peak needs SunE Alamosa – Single Axis Tracking (7.1AC) August 4, 2011 July 1, 2011

FEBRUARY 2011 Week 1 Week 2 Week 3 Week 4 Wind Energy Integration Wind variability is over a longer time frame – tens of minutes to hours. Production peaks at all times of the day – not just at night. Wind has a low capacity value over summer peak periods PSCo Wind Output (MWh)

Wind Energy Integration In real-time, more wind means fewer dispatchable resources are on-line to manage the variability. Additionally, wind competes with traditional baseload facilities for system bandwidth. Occasional wind energy curtailment is required to maintain system reliability.

Is Storage the Solution?

Charging Fuel (% of Time) vs. CO 2 Cost System with Coal Base Load

Relative Value of Storage PSCo System (Coal Baseload) High Value Low Value

Wind Integration Costs: Uncertainty and Variability Public Service Company of Colorado 2 GW and 3 GW Wind Integration Cost Study. August 19, Pg 24. Energy Storage Reduces Wind Integration Costs Upgrade Cabin Creek  Increase upper pond  Higher nameplate capacity  Increase turnaround efficiency Additional Storage Resource  2 nd Cabin Creek

Conclusion Currently, PSCo system uses Cabin Creek and gas storage as buffers System doesn’t have sufficient wind curtailment exposure that can be cured by energy storage in the near-term Storage capital costs above arbitrage value BUT…as more renewable energy is integrated into the system, energy storage provides more value. We will continue to evaluate role and economics of storage.

Near-Term Mitigation Strategy Lower the Floor Lower plant minimum capacity Decommit - Reevaluate “must run” status Decommission baseload (CACJA) Upgrade Existing Storage Facilities (ie Cabin Creek) Increase upper pond capacity Move black start capability to diesel generators Enable Wind to Participate in System Balancing Implement set point control at wind plants (“optimal” curtailment) Better forecasts = Better planning Long-Term Solutions… Add New Storage Facilities Dispatchable Loads

A Big Picture View of Wind Curtailment

System Perspective (2013)

Wind

System Perspective (2013)

Wind Coal Gas

Visualization of How Off-Peak Load Helps Mitigate Wind Curtailment

System Bottoming Effects (No Storage) System with Coal Baseload

System Bottoming Effects (With Storage) System with Coal Baseload

Xcel Energy Wind and Solar Future

Xcel Energy Wind Farms Today PSCo – 1735MW SPS – 736MW NSP-MN – 1591MW Xcel Energy – 4062MW Xcel Energy manages output from 4.1GW (2972 turbines) of wind energy across the three operating companies (NSP-MN, PSCo, SPS) and seven states (CO, TX, NM, WY, MN, SD, ND) Forecast by end of 2011 Q3

Xcel Energy Wind Generation Growth Xcel Energy has been the nation’s largest wind power provider for seven consecutive years.* Current resource plans have 5GW installed by PSCo will be at nearly 2GW by the end of Installed Wind Capacity (MW) *American Wind Energy Association (AWEA); US Wind Industry Annual Market Report 2010

Colorado Solar Generation Growth HB creates minimum retail distributed generation* requirement Requirement is largely met through Solar Rewards program (current 99.4MW) Solar Gardens legislation (HB ) allows for community solar facilities (<2MW) to qualify as well. Distributed Solar PV Capacity *Retail Distributed Generation - Must be renewable energy (wind, solar, bilmass, hydroelectric) installed on the distribution system. There are further requirments regarding the how much can be residential vs commercial/industrial, amount installed by year, etc

Why Forecasting?

Obligations and Resources Today Net Load – Load less variable output generation. Traditional Utility Paradigm Mid-Merit Baseload Peakers Renewables PSCo - April 26, 2010 to May 2, 2010

Loads and Resources Today Loads and Resources are forecast every working day for operational planning purposes. This is called day-ahead commitment. PSCo - April 26, 2010 to May 2, 2010 Actual – Thick Line; Forecast – Thin Line Forecasts of loads, variable energy generation, and unit availability are inputs to the operational planning process. Typically 18 to 42 hours ahead, but as much as five days. Wind Loads

Loads and Resources Today Uncertainty is driven by the wind portfolio. PSCo - April 26, 2010 to May 2, 2010 Mean Absolute Error Load = 86MW Wind = 245MW Net Load = 240MW

Renewable Energy Integration Solutions

Loads and Resources – A New Paradigm High Penetration Portfolio Balance Portfolio Variable Traditional Utility Paradigm (w/ some RE) Mid-Merit Baseload Peakers Renewables RE Forecasts are nice but don’t affect operations RE is must-take Fossil-based facilities operate to a define duty New facilities chosen for least-cost energy Storage facilities dispatched to peak shave RE Forecasts drive operational decisions RE is dispatchable Fossil-based facilities are modified for flexible duty New facilities chosen for least-cost energy w/ flexible optionality Storage facilities dispatched for RE/load shifting