Information Backbone for a Sustainable Electric Power System Information technology will profoundly transform the planning and operation of the power grid … Rob Pratt Pacific Northwest National Laboratory University of Washington Seminar December 2007 1
Today’s Discussion Introduction to the Grid. Wise program at PNNL Market-based demand response for managing peak loads Demand response for reliability – Grid Friendly™ Appliances Grid. LAB-D simulation of smart grid technologies and operations Interoperability for integrating demand response and other distributed resources Smart grid’s role in sustainable electric power system Impact of high penetration of plug-in hybrid electric vehicles Electric Infrastructure Operations Center and advanced transmission R&D 2
Communicate – With Whom? About What? Customer Perspective ofretrofit voltage, a Grid. Wise World capacity, availaudit results, Gen, T, & D Suppliers ability, price, forecast, contract terms , DG incentives opportunities, designs, costs, terms & conditions DG/storage Distribution status Linemen grid status level, Energy Service end-use power/ Co. s, Vendors, rations Utility Programs Aggregators billing, info access, attractive contracts, approvals, occupancy, performance power requirements, forecasts, status, curtailment Emergency Operations Customer Appliances, Equipment, Processes 3
Fully-Engaging Demand Response in Grid Operations Rob Pratt PNNL Energy Sciences & Technology Directorate
Value of Demand Elasticity: Lower Peak Demand & Stabilize Prices Demand (elastic) Price ($/MW) Price Demand (inelastic) Price, mitigated Supply Quantity (MW) 5
How Does Managing Peak Electrical Demand Save Money? Hourly Loads as Fraction of Peak, Sorted from Highest to Lowest 90% generation 75% 5% = ~400 hrs/yr distribution 5% (8, 760 hrs) 25% of distribution & 10% of generation assets (transmission is similar), worth of 100 s of billions of dollars, are needed less than 400 hrs/year! 6
Pacific NW Grid. Wise™ Testbed Projects Olympic Peninsula Two Current Projects: Olympic Peninsula Grid. Wise Demonstration Grid Friendly™ Appliance Demonstration Yakima Gresham 7
Olympic Peninsula Project Team U. S. DOE/OEDER Whirlpool Corporation Bonneville Power Administration IBM Portland General Electric Invensys Controls Pacifi. Corp Clallam County PUD Preston Michie and Associates City of Port Angeles Dr. Lynne Kiesling Montana Tech $2. 0 M project funding over two years through PNNL, technology innovator, project manager $75 K project funding, Non-wires Program resources, in-kind labor $63 K project funding over three years, utility site host, inkind meter installation labor $50 K project funding, utility site host, in-kind recruitment labor Energy pricing consultant and valuation analysis Economic experiment design consultant Manufacturer of Sears Kenmore™ HE 2 dryer, vendor, inkind research labor Provider of communications technologies, in-kind application development labor and Web. Sphere™ software provider Residential communication and control equipment, vendor of Good. Watts™ system Utility site host, in-kind meter installation labor, in-kind recruitment labor Student labor collaboration 8
Olympic Peninsula Demonstration IBM Invensys $/k. Wh Market ancillary services distribution congestion $ transmission congestion MW 0 Clallam PUD & Port Angeles n = 120, 0. 5 MW DR Johnson Controls 6 12 wholesale cost 18 24 Clallam County Internet broadband PUD Water communications Supply District Sequim Marine 0. 2 MW DR Sciences Lab 0. 3 MW DR Johnson 0. 5 MW DG Controls 9
Economic Experiment Approach Establish, offer and compare three retail contract types: fixed-price, time-of-use, real-time (5 -minute) price Provide automation so residents can configure space conditioning and water heating for their relative comfort vs. economy preferences Establish a shadow market and compensate residents for the degree to which they respond to price signals (i. e. , energy cost savings) Residents receive “grant” with which to pay shadow market electric bill, and keep the balance (they pay their regular electric bill as usual) 10
RTP Control with “Virtual Thermostat” Price (Cooling Example) User sets: Tdesired, comfort vs. economy setting These imply: Tmax, Tmin, k k Pavg + k Pbid Pavg Pclear Pavg - k Temperature Tmin Tset Tdesired Tcurrent Tmax Smaller k: lower comfort, higher demand response, higher savings Larger k: higher comfort, lower demand response, lower savings 11
Managing a Transmission or Distribution Constraint with Energy Prices in a Series of Peak Load Events DG required above feeder limit Market failed to meet demand for one 5 -min. interval in 3 -day cold snap 12
. Grid Friendly™ Appliances (GFAs) Help Keep the Lights On! Loads and Reserves on a Typical U. S. Peak Day Industrial 28% Commercial 29% Resident. (non. GFA) 12% Residential (GFA*) 18% GFA* potential exceeds US operating reserve requirements! Operating reserves 13% * GFA for: heat, AC, HW, refrigerators, freezers Grid Friendly Appliances sense grid frequency excursions & control region’s appliances to act as spinning reserve – No communications required! 13
Stabilization Potential from Frequency Responsive Load Bus 25 frequency @ t = 1 sec: loads +5% @ t = 40 sec: loads +15% from Trudnowski et al. IEEE PES. 2005. (http: //gridwise. pnl. gov/docs/pnnlsa 44073. pdf) 14
Grid Friendly™ Appliance Demonstration Autonomously detects under-frequency events, sheds load for up to a few minutes 150 new Whirlpool clothes dryers, 50 retrofitted water heaters No one noticed in hundreds of curtailment events! An ancillary service that can displace spinning reserves and increase reliability Reacts within 1/2 second Delays & randomizes service restoration to avoid shocking the grid, eases cold load pickup after an outage Low cost: no communications required Vast, inexpensive grid “safety net” “When the inevitable occurs … people get stuck in elevators and high-value uses of power are shut off along with all the lowest priority uses of energy. It's the meat-ax approach to interrupting power flows. ” Dr. Vernon Smith, 2002 Nobel prize Winner, Economics 15
All GFAs Respond Within Seconds of Event 16
What we’ve learned Important insights from our Northwest demo project Residential customers will respond to 5 -min. real-time prices (RTP) if provided opportunity to save 10%, and with technology that makes it simple to automate their preferred response Demand response + distributed generation provided sustained, local distribution peak load reduction (in addition to wholesale market benefits) l l able to cap net demand at an arbitrary level, ~15% less than the normal annual peak; will respond for several sequential days real capital cost savings when a $5 -10 M substation can be deferred or downsized A portfolio of RTP, TOU, & fixed rates may be optimal l l RTP focuses demand response just when it is needed Time-of-use (TOU) resulted in efficiency gains RTP and TOU can be implemented without an actual rate change, and with “no losers” compared to flat rates, by debiting against an up-front credit 17
What we’ve learned (cont. ) Important insights from our Northwest demo project Using demand to provide short-term regulation is a simple, inexpensive byproduct of an RTP system (i. e. , can help manage high penetration of intermittent wind generation) l l can easily synchronize thermostatically controlled loads to follow grid’s need for regulation over the short term (minutes) excursions from desired setpoints are very small minimal if any discomfort means costs to buy response are very low likely at far lower costs than power plants charge to ramp up/down Autonomous under-frequency/under-voltage load shedding from Grid Friendly™ appliances was reliable and noticed by users 18
Grid. Lab-D – a Collaborative, Open Source Environment for Simulating the Smart Grid Detailed, simultaneous simulation of power flow, end use, and market functions and their interactions Evaluate the potential of new technologies and distribution system operational strategies to: l l l save capital costs improve reliability provide other benefits Craft and refine the characteristics of technologies and operational strategies to provide maximum benefit at lowest cost Understand quantify the synergies of deploying a broad range of smart grid technologies Avoid unintended consequences that can result from utilizing distributed control systems Predict and evaluate results from deployment projects 19 19
Grid. Wise Architecture Council Linking Islands of Enterprise Integration the Building+Grid Enterprise? ? ? the Grid Enterprise the Building Enterprise OR with a minimal Transactive Interface Maintenance HVAC Control Planning Operations Contract: Accounting commodity, price, data, terms and conditions… Billing 20
Smart grid technology provides the critical information backbone for a sustainable electric system Deliver peak load demand reduction Increase reliability benefits Squeeze more power through existing assets Deliver efficiency in addition to peak demand savings Measure and validate carbon offsets Integrate renewables l Demand response support high wind penetration l Reward solar photovoltaic's for peak availability 21
Load Duration Curve and Carbon Intensity of Marginal Generation 2. 0 Carbon Intensity of Marginal Plant Load Duration Curve Peaking Combustion Turbine Intermediate Combined Cycle 1. 5 Baseload Coal Steam CO 2 (lb/k. Wh) 1. 0 0. 5 0. 0 22
Carbon “Supply Curve” Suggests Massive Efficiency Efforts will Come Early 23
Potential Impacts of High Penetration of Plug-in Hybrid Vehicles (PHEVs) on the U. S. Power Grid* * PNNL study for DOE Office of Electricity The idle capacity of today’s U. S. grid could supply 73% of the energy needs of today’s cars, SUVs, pickup trucks, and vans… without adding generation or transmission if vehicles are managed to charge off peak 52% Source: EIA, Annual Energy Review 2005 73% electric (158 million vehicles) Potential to displace 52% of net oil imports (6. 7 MMbpd) More sales + same infrastructure = downward pressure on rates Reduces CO 2 emissions by 27% Emissions move from tailpipes to smokestacks (and base load plants) … cheaper to clean up Introduces vast electricity storage potential for the grid 24
Fundamental Approach 1: Determine Available Marginal Generation normalized electric system load ` 1. 20 1. 00 Regional peak day load profile (NERC 2002) Dispatched generation to meet load (matches EIA 2002 annual totals) Available generation (EIA 2002) No additional generation from existing: • Nuclear • Hydro • Renewables Regional avg. peak season load profile (NERC 2002) 0. 80 0. 60 0. 40 0. 20 Assumptions: Combined cycle Fossil steam Coal Combustion Turbine Renewables Hydro Nuclear • Combustion Turbines (peaking plants) 0. 00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 hour of day 25
Fundamental Approach 2: “Fill the Valley” in the Load Shape Assumption: normalized electric system load ` 1. 20 1. 00 0. 80 0. 60 0. 40 0. 20 0. 00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 hour of day Additional valley-filling generation constrained to lesser of: • Available marginal generation @ 85% capacity factor • Peak load Combined cycle Fossil steam Coal Combustion Turbine Renewables Hydro Nuclear 26
Coal, Nat. Gas Power Plants Fill the Valley nighttime valley-filling daytime valley-filling Summary normalized electric system load ` 1. 20 Determine size of valley in MWh 1. 00 l Combined cycle l 0. 80 Fossil steam 0. 60 Coal 0. 40 No marginal added generation in valley from: l 0. 20 l l 0. 00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 hour of day Floor: average day in the peak season Ceiling: lesser of available marginal generation @ 85% or peak load l Hydro All other renewables Nuclear Peaking plants 27
Analysis by NERC Region* 105% 57% 135% 79% MAPP 18% 10% 127% NWP 23% 15% CNV 78% 104% 46% 61% MAIN 66% 39% 100% 45% NPCC(US) 73% AZN & RMP 80% ECAR SPP 52% 31% SERC 73% 86% 49% 57% 34% ERCOT Nighttime Charging Only (hrs 18 – 6) Daytime + Nighttime Charging (0 – 24 hrs) Summary u Midwest: support almost the entire LDV fleet u East: somewhat smaller potential u West: supports fewer vehicles % figures denote the percentage of LDV fleet supported by idle electric capacity 28
Regional Emissions Impacts (Well-to-Wheel*) with Today’s Generation Mix * Argonne National Laboratory’s GREET well-to-wheel model Existing coal plants break even on greenhouse gases Nationally, greenhouse gases reduced 27% despite increased reliance on coal Plant mix for valley fill Greenhouse gases Particulates SOx Urban: VOCs CO NOx Particulates SOx from vehicles doubles: cap-and-trade will require investment in cleaner plants Moving emissions from tailpipes to smokestacks: l l solves an intractable problem for CO 2 capture improves cost effectiveness for other emissions Urban air quality emissions greatly reduced: VOCs/CO/NOx > 90% SOx = 80% Particulates = 40% 29
Information: The Virtual Electric Infrastructure FACT: CHOICE: In the next 20 years, the U. S. will spend $450 B on electric infrastructure, just to meet load growth. Perpetuate a 20 th Century solution OR Invest in a 21 st Century system saving ratepayers $80 B while increasing reliability and flexibility. : Revealing Values + Communications + Advanced Controls ≡ Electric infrastructure The choice is easy because… $ bits << $ iron
Energy Sciences and Technology Directorate Electricity Infrastructure Operations Center Fully capable grid control center for training and backup Live data resources from actual grid operations nationwide State-of-the-art grid operation & modeling tools (AREVA T&D partner, $3 M in-kind software) Supporting network hardware and computation capabilities Access to advanced computers and high speed networks Unique platform to research, develop & test next generation tools and concepts for operating the energy infrastructure Electricity Infrastructure Operations Initiative 31
Energy Sciences and Technology Directorate Electricity Infrastructure Operations Center An industry-native working environment: understand quantify current operations as a baseline point-of-departure conduct R&D for new control and operation technologies evaluate and quantify their benefits transfer technology to … conduct advanced training for … and obtain feedback from … power system operators understand human factors and develop visualization techniques that enhance situational awareness. Electricity Infrastructure Operations Initiative 32
High-Speed Grid Computing Research Static States Point-of-departure: Static State Estimation Data point Once the cascade began, the 2003 blackout swept from Ohio to NY in nine seconds! Data collection cycle Actual conditions Measured data ± error pu Curtaining ting mc ut so& pom & satec t ti-es im es 2 i 4 asemate k-4 m-n min s ge-te ta ng sn t asialyakes 2 eisolvita R n s t sis esolv R cy ennaly ontncy a geing c contin Operators had no way to see imminent instability! Presumably quasi-steady state Resolved state estimate 4 sec 8 sec 12 sec 2 -4 min Electricity Infrastructure Operations Initiative Time 33
High-Speed Grid Computing Research (cont. ) Goal: Dynamic Situational Awareness at Data Cycle Speed • Dynamic state estimation Dynamic States • Dynamic stability & contingency analysis: off-line → real-time • Advanced scientific computing architectures & algorithms Resolved dynamic state estimates & parameters Data collection cycle • Engage DOE Office of Science More throughput and more reliability means $$$ • Minutes → seconds means 102 speed-up • New phasor data cycle is 1/30 sec means 104 Curtain Look-ahead simulation Calibrated dynamic model 4 sec 8 sec 12 sec 1 min Electricity Infrastructure Operations Initiative Look-ahead provides time to react & stabilize grid Time 34
Watershed Management/Hydro Operations Research Goal: Scientific basis for real-time, joint optimization of hydro & fish • Real-time assimilation of remotely sensed data • Multi-scale analysis • Real-time hydraulic modeling of fish conditions: static→dynamic constraints • 7000 sub-basins down to 10 m • Ensemble climate & streamflow forecasting • Multi-objective optimization supported by highspeed computing • Synexus Global partnering with Rivers Mixing at Confluence of Snake & Clearwatertheir existing BPA model HUC 5 HUC 6 at Confluence of Snake & Clearwater Rivers m 10 Mixing Electricity Infrastructure Operations Initiative 35
Rob Pratt, Program Manager Pacific Northwest National Laboratory robert. pratt@pnl. gov 509 -375 -3648 U. S. Department of Energy Office of Electric Delivery and Energy Relibility http: //www. oe. energy. gov/ Electric Distribution/Grid. Wise http: //www. electricdistribution. ctc. com/ Grid. Wise Architecture Council http: //www. gridwiseac. org/ Grid. Wise at PNNL http: //gridwise. pnl. gov/
Скачать презентацию Information Backbone for a Sustainable Electric Power System
05ec55d2b2eed25de901e696e6303d8a.ppt