Demand Response Forecasting, Measurement

1
Demand ResponseForecasting, Measurement
VerificationMethods/Challenges/Considerations

• National Town Meeting on Demand Response
• Washington, DC - June 2 3, 2008
• Mark Williamson
• Manager, Load Research, DTE Energy

2
What is a Baseline?
Baseline The amount of energy the customer
would have consumed absent request to reduce
3
Forecasting Measurement
Class Characteristics
Likely Methods

• Sampling
• Regression
• Engineering Models
• By Class

• Baseline (CBL) estimate
• Average of Days
• Interval Metering
• By Customer

4
Measurement of Actual DR Event Direct Load
Control w/Samples
5
Not All Loads Are Alike Key Considerations to
Baselines
One Size Does Not Fit All

• Several factors require consideration when
  estimating or measuring Demand Response impacts
  which may dictate utilizing multiple methods.
• Weather vs. Non-Weather sensitive
  customers/classes
• Type of customer (number of shifts, hours or
  operation)
• Seasonality
• Historical data
• Day of event adjustment method
• Geographic region
• Time Impact / Time Zone
• Slope of load curve day of event

Slope impacts day of adjustment
High rise office building, monitored HVAC,
weather sensitive
Manufacturing, 3 Shift, or non-weather sensitive
6
Are Current Methods Acceptable?
7
Challenges and Move to M V Standards

• Challenges
• Leverage existing work adjusting for key
  learning's
• Improve quality of baseline estimate
• Day of event baseline adjustment
• Incorporate weather sensitivity in original CBL
• Differentiate weather vs. non-weather sensitive
  loads
• Minimizing number of methods
• Keep it simple if possible
• M V Standards
• Continue effort to develop national standards
• Recognize concerns of all participants
• Account for jurisdictional issues
• Develop standard controls reporting

8
Baselines (CBL) -What should be considered?
Based on number of day average Do not Consider
type of business Weather Season Day of Week