New Developments in Demand Response

1
New Developments in Demand Response

• April 28, 2005
• David S. Watson
• Lawrence Berkeley National Lab
• watson_at_lbl.gov
• 510.486.6255

2
Presentation Overview

• Demand Response Overview
• DR Policy
• Value of Demand Response
• Programs and Tariffs
• Communications, Controls Monitoring
• Case Studies
• Consumer Institutional Behavior
• Auto-DR System Characteristics
• Shed Strategies
• State-of-the-Art DR Features
• The Future of DR in California

3
Lawrence Berkeley National Lab

• Discovered 16 Elements (e.g., Californium)
• 10 Nobel prizes
• 200 Building Science Researchers
• Didnt buy the Ford Mustang with the Lab credit
  card (that was Livermore!)

4
DR Research Context

• Deregulation in CA in 1996 (AB 1890)
• California electricity crisis of 2000-2001
• Aging fossil fuel plants (pollution)
• Flaws in deregulation (AB 1890)
• Disconnect between wholesale and retail prices
• CA ISO (32) Stage-1 (lt7) and (17) Stage-2
  alerts (lt5 reserves)
• Statewide rotating blackouts
• 3 days in January, 2 in March, 2 in May 2001
• Billions in lost revenues

5
DR Definition

• DR Definition action to reduce load when
• Contingencies occur that threaten supply-demand
  balance
• Market conditions occur that raise supply costs
• peak-load reductions different from efficiency,
  transient vs. permanent

6
Why do Demand Response?

• Need DR because
• Load on grid is constant flux
• Generating capacity not always adequate for peaks
• Transmission capacity not always adequate for
  peaks
• DR less than ½ the cost of adding capacity (per
  kW)

7
Long-Term Options

• Build more plants (supply approach)
• Replace aging (dirty) fossil fuel plants
• Add peaker plants (costly, NIMBY)
• Add transmission lines (costly, NIMBY)
• Use load as a resource (demand approach)
• Minimal DR mitigates major imbalances (3)
• There is a large potential load resource

8
DR PolicyWhat are we trying to achieve?

• Joint proceeding CEC and CPUC (R.02-06-001)
• Early goal for price sensitive demand response
  5 of peak by 2007
• Residential Default CPP
• Small Commercial (lt 200 kW) Default CPP
• Medium Commercial (lt 999 kW) Default CPP
• Large CI (gt 1 MW) 2-part RTP

9
How much effect can DR have?
10
Shed Control Signals (1 of 2) - Overview

• Most programs managed by utility or ISO
• Triggered by weather or low reserves of
  electricity
• Traditionally
• Dispatch sheds via phone calls, pagers, e-mail
• Sheds implemented manually
• Slow, labor intensive
• The future
• Dispatch sheds via wireless, powerline, Internet
• Automated sheds without human intervention
• Opt out capability

11
Shed Control Signals (2 of 2) - Communications

• Load shedding device is controlled remotely via
• Wireless
• 1-way proprietary radio based systems
• 1-way and 2-way pager based systems
• Powerline
• Control signals are sent over the existing power
  lines
• Internet
• Control signals are sent over the public Internet
  and/or private networks

12
Shed Measurement (1 of 4)

• Energy Management and Control System (EMCS)
• Primary function HVAC scheduling control
• Usually dont have meters on system
• Data logging often problematic
• Energy Information System (EIS)
• Designed for Energy
• Often Web based
• No control

13
Case Study - Auto-DR tests were conducted at 18
sites in 2003 2004
Including Albertsons Supermarket, Oakland
BofA Bank Office Building, Concord GSA
Govt Office Building, Oakland UCSB College
Campus Library, SB Echelon Office Building,
San Jose
14
Auto-DR System Description

• LBNL published a fictional, variable price signal
  (/kWh) using XML/Web Services
• Commercial facilities automatically checked the
  price
• Electric loads automatically shed upon a rise in
  price
• No human intervention

15
Geographic Distribution

• All facilities were in California
• Auto-DR infrastructure nationwide

16
Auto-DR System Definitions

• Price Server Common source of current price
  info
• Polling Client Software that polls the server
  to get price
• Business Logic Decides EMCS actions based on
  price

17
Auto-DR Systems Polling Clients

• Polling client is the software application that
  checks (polls) Price server to get the latest
  price. This is known as pull architecture
• The polling client (software) resides on a
  computer managed by each site
• Each client polled the server each minute
• Client passes price to business logic software.
  Determines if action is required in EMCS devices
  

18
Auto-DR System Communications

• LBNL defined price schedule
• Price published on XML server
• Clients request price from server every minute
  send shed commands
• EMCS carries out shed

19
Auto-DR Systems Test Sites - Circa 1999
20
Auto-DR Systems test Sites - Circa 2002 (After
CEC Program)
21
Auto-DR Systems Test Sites Nov. 2003 Auto-DR
Test
22
Auto-DR System Fictional Price Signal Nov. 19,
2003

23
Results Site A Shed (lights and anti-sweat
heaters)
24
Effect of Lighting Load Shed on Store Lighting
Reduced Lighting during Shed
Full Lighting
25
2004 Re-Test, 90 F Day
Max of 1450 kW Shed
26
Load Shapes from Office Buildings
27
Results from 2nd Test in 2004 Power Savings
(W/ft2)
28
Results 2004 1.5 deg. F Temperature Rise at GSA
(39 zone average)
29
Consumer and Institutional Behavior

• Multi-tenant leased office buildings have trouble
  participating in DR
• Should consider lease terms allowing service
  deviations based Government declared grid
  emergencies
• Owner occupied and government buildings have
  highest levels of participation
• Virtually no comfort complains in any the LBNL DR
  tests

30
Issues - Chicken Cooker added to Lighting Circuit
(undocumented)
31
Demand Shifting with Thermal Mass (Pre-Cooling)
(1 of 2)

• Goal - understand demand shifting with building
  mass develop optimal control
• RD Team So. Cal. Edison, Purdue,
• UC Berkeley, LBNL
• Current Work commercial building field studies
  preliminary simulation study

Concrete Floor
Thermal Capacity 3 Watts-Hours/ft3 - F
32
Pre-Cooling Results from 2003 (2 of 2)
33
Auto-DR Systems Shed Control
34
Identifying DR HVAC Strategies (1 of 2)
35
(No Transcript)
36
Auto-DR System Findings What is
State-of-the-Art ?(1 of 5)

• The best system meets or exceeds the
  requirements at the lowest installed cost
•  Flexible Designs for the Future
• As future-proof as possible
• Features
• Customer defined business logic
• Customer warnings, alerts local overrides
• Real-time, two-way verification of shed
• High security (shed) availability

37
Auto-DR System Findings What is
State-of-the-Art ?(2 of 5)

• Leverage Trends in Technology
•  Internet and private WANs are ideal platforms
  for Auto-DR controls and communications
•  Performance of IT equipment (e.g. routers,
  firewalls, etc.) continues to improve and prices
  continue to drop

38
Auto-DR System Findings What is
State-of-the-Art ?(3 of 5)

• Enterprise, EMCS EIS Integration
•  EISs and EMCSs should share the same control
  network, database and HMI.
•  EMCS/EISs networks should be tightly integrated
  with enterprise networks
•  The most robust and least costly systems should
  have no more than 1 enterprise protocol and 1
  control protocol

39
Auto-DR System Findings What is
State-of-the-Art ?(4 of 5)

• Open protocol standards should be used
• Allows the greatest flexibility at the lowest
  cost
• Truly open systems are interoperable
• XML alone is not a standard for building and
  energy data. OPC-XML, oBIX, and others are
  developing XML standards of this type

40
Auto-DR System Findings What is
State-of-the-Art ?(5 of 5)

• Shed Strategies
• Should be designed to minimize discomfort,
  inconvenience, and loss of revenue
• High granularity closed loop control are less
  likely to negatively impact building occupants
  for a given demand shed
• In addition to HVAC control strategies, lighting
  and other loads should be considered for sheds as
  well

41
Next Phase Auto-DR Auto-CPP Tests Summer 2005

• Currently recruiting commercial sites !
• Participating sites will get assistance preparing
  for future dynamic pricing programs.
• Site requirements
• Located in PGE territory
• Have or want remote telemetry
• Cooperative staff
• Willing to shed some electric loads for several
  automated tests
• Contact Dave Watson (LBNL) watson_at_lbl.gov for
  more info

42
Demand Response Research Center

• Objective to develop, prioritize, conduct, and
  disseminate multi-institutional research to
  facilitate DR
• Scope technologies, policies, programs,
  strategies and practices, emphasizing a market
  connection
• Method Partners Planning Committee, Annual RD
  Plan
• Stakeholders

• State policy makers
• Researchers
• Information metering system developers
• Aggregators
• Program implementers

• Utilities
• Industry trade associations
• Building owners, engineers operators
• Building equipment manufacturers
• Other end-use customers

43
Further information

• Dave Watson
• 510 486-5562
• watson_at_lbl.gov
• LBNL Demand Response Research Center
• (mngr. Mary Ann Piette)
• mapiette_at_lbl.gov
• http//drrc.lbl.gov/drrc.html
• Peng Xu, pxu_at_lbl.gov
• Pre-cooling Research
• Chuck Goldman, CAGoldman_at_lbl.gov
• DR Policy Research

• LBNL Building Technologies
• http//eetd.lbl.gov/BT.html
• CEC Integrated Energy Policy Report (incl. Demand
  Response goals)
• http//www.energy.ca.gov
• /2004_policy_update/