FAST v3

538 views

Published on

0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
538
On SlideShare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
3
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

FAST v3

  1. 1. cs/ee 146 Fri: Smart Grid<br /><ul><li>Smart Grid intro
  2. 2. Course organization</li></ul>Steven Low<br />CMS, Caltech<br />Winter 2010<br />
  3. 3. cs/ee 146<br />W<br />F<br />
  4. 4. cs/ee 146<br />Wed<br />Fri<br />
  5. 5. Gencos<br />Transcos<br />ISO/RTO<br />LSE<br />(Utilities)<br />
  6. 6. Where does electricity come from<br />generator<br />generator<br />transmission<br />Transcos<br />substation<br />Transmission: <br />enerator to <br />distribution<br />Transmission: from generator to substation, long distance, high voltage<br />Distribution:from substation to customer, shorter distance, lower voltage<br />
  7. 7. Outline<br /><ul><li>Snapshot of US power grid
  8. 8. Distribution
  9. 9. Generation
  10. 10. Transmission
  11. 11. Where can we help?
  12. 12. cs/ee 146 projects</li></li></ul><li>US power distribution<br />Nation’s electricity bill: $247B<br />131M customers<br />Average price: 7c/KWh<br />3,100 electric utilities<br />213 private utilities provide ~73% of customers<br />2,000 public (state & local) utilities provide ~15% of customers<br />865 distribution coops and 65 generation & transmission coops provide ~12% customers<br />Source: DoE, GRID 2030, 2003<br />
  13. 13. US power distribution<br />Source: <br />DoE, 2008<br />Smart Grid Intro<br />
  14. 14. Problem with peak<br /><ul><li>National load factor: 55%
  15. 15. 10% of generation and 25% of distribution facilities are used less than 400 hrs per year, i.e. ~5% of time
  16. 16. Existing power plants can provide 73% of light vehicles
  17. 17. If they are plug-ins that recharge at night
  18. 18. Will reduce oil consumption by 6.2M barrels a day</li></ul>Source: DoE, Smart Grid Intro, 2008<br />
  19. 19. US power generation<br />9,200 power plants with 1TW capacity<br />World electricity usage: 1.9 TW / 297 Wpc(2005);<br />Electricity usage<br /><ul><li>US: 436 GW / 1460 Wpc(2005);
  20. 20. China: 326 GW / 248 Wpc(2006);
  21. 21. EU: 322 GW / 700 Wpc(2004)</li></ul>Source: DoE, Smart Grid Intro, 2008<br />
  22. 22. US power generation<br /><ul><li>Central power plants </li></ul>33% thermal efficiency, unchanged since 1960s<br />Central power generation cannot recycle heat<br />Distributed energy facilities <br />~5,600 facilities, 6% of power capacity (2001)<br />65% - 90% efficiency<br />Combine heat & power generation<br />Source: DoE, GRID 2030, 2003<br />
  23. 23. Future trend<br />Source: DoE, Smart Grid Intro, 2008<br />
  24. 24. Generate more than power...<br />US CO2 emission<br />Elect generation: 40%<br />Transportation: 20%<br />Source: DoE, Smart Grid Intro, 2008<br />
  25. 25. US Energy Flow -1999Net Primary Resource Consumption 102 Exajoules<br />
  26. 26. US Energy Flow -1999Net Primary Resource Consumption 102 Exajoules<br />
  27. 27. Outline<br /><ul><li>Snapshot of US power grid
  28. 28. Distribution
  29. 29. Generation
  30. 30. Transmission
  31. 31. Where can we help?
  32. 32. cs/ee 146 projects</li></li></ul><li>US power transmission<br />300,000 miles of transmission & distribution lines<br />157K miles of high voltage (>230kV)<br />Trans & distr losses: 9.5% (2001)<br />Trans & distr losses: 5% (1970)<br />99.97% reliable<br />Yet, outages cost $150B/yr<br />Northeast blackout of 2003: $6B loss<br />5 massive blackout in the last 40 yrs<br />3 of which occurred in the last 9 yrs<br />Source: DoE, Smart Grid Intro, 2008<br />
  33. 33. How much is reliability worth?<br />Source: DoE, Smart Grid Intro, 2008<br />
  34. 34. US power grid<br />Source: Roger Anderson, <br />Columbia University 2004 <br />
  35. 35. US power transmission<br />Old infrastructure<br />Average generating station build in 1960s<br />Average age of substation transformer: 42 yrs (expected life span: 40 yrs)<br />Before PCs and Internet!<br />Since 1982, peak demand outgrows transmission capacity by 25%/yr<br />From 1988-98, electricity demand rose by ~30%, while transmission capacity grew by ~15%<br />Annual investment in transmission capacity has dropped 50% since 1975 [Wu et al, Proc IEEE, 2005]<br />Source: DoE, Smart Grid Intro, 2008<br />
  36. 36.
  37. 37. Difficulties in transmission expansion<br /><ul><li>Private info after deregulation
  38. 38. Gencos, transcos, utilities
  39. 39. Different timescale
  40. 40. Transmission projects: 5-10 yrs
  41. 41. Generation projects: ~2 yrs
  42. 42. Uncertainty in matching supply & demand
  43. 43. Renewable sources, demand response
  44. 44. Substitution between trans capacity, gen capacity, demand mgt
  45. 45. Uncertainty in RoI& regulations, environmental opposition</li></li></ul><li>Some benefits of smart grid<br />5% higher grid efficiency = 53M cars<br />Real-time dynamic visibility of power system<br />Now: measurements at 2-4 s timescale offers steady-state behavior<br />Future: measurement at ms timescale offers dynamic behavior<br />But: lack theory on how to control<br />Source: DoE, Smart Grid Intro, 2008<br />
  46. 46. Enabling technologies<br />AMI (Advanced Metering Infrastructure)<br />Enable demand response<br />Promote user participation<br />PMU (Phasor Measurement Unit)<br />Real-time, global, synchronized measurement <br />At ms timescale<br />ICT integration with grid<br />High speed WAN allows real-time and global monitoring at control centers<br />High performance computing allows faster control decisions<br />Integrating renewables & distributed sources<br />
  47. 47. Conclusions<br />Grid connects generation to load<br />Gridmust become more efficient<br />Demand and generation will continue to outpace transmission capacity<br />Grid must integrate distributed and renewablesources<br />Sustainability<br />Grid must promote demandresponse<br />Cutting peak demand has a large impact<br />Integrate and exploit EV<br />
  48. 48. Outline<br /><ul><li>Snapshot of US power grid
  49. 49. Distribution
  50. 50. Generation
  51. 51. Transmission
  52. 52. Where can we help?
  53. 53. cs/ee 146 projects</li></li></ul><li>Where we can help<br />Distributed sense & respond in SG<br />Integration of solar and wind power<br />Pricing and optimization of demand response<br />Architecture and protocols to implement these algorithms<br /><ul><li>These problems are interconnected
  54. 54. Each requires engineering+economics</li></li></ul><li>Current grid control<br />First line of defense<br />Protective relay system<br />Fault detected by current or potential transformer<br />Triggers relay to trip appropriate circuit breakers to isolate fault<br />Timescale: ~10ms (a couple cycles)<br />Based on local info<br />
  55. 55. Current grid control<br />Second line of defense<br />Control center EMS<br />Substation RTUs sample power measurements<br />SCADA sends measurements to CC every 2-4s<br />State estimator computes a snapshot of system equilibrium<br />N-1 contingency simulation of dynamic behavior<br />Preventive control: change system operating conditions before a fault to ensure the system can withstand the fault<br />Timescale: 2-4s<br />Based on global info<br />
  56. 56. Current grid control: issues<br />Different timescales drive independent actions by relay system and CC<br />This gap has led to missed opportunity in preventing cascading outages<br />E.g., NA blackout & Italian blackout of 2003<br />No analytical tools for emergency control<br />No dynamic system model for stability analysis after fault<br />Rely on simulation: too slow for real-time control<br />Necessitates conservative open-loop control<br />
  57. 57. Solution: grid+ICT<br />Both issues due to ICT limitation<br />Data acquisition system (SCADA)<br />CC computational power<br />... that can be drastically improved<br />More computational power in CC<br />Faster SCADA communications over WAN<br />GPS-based PMUs provide real-time and synchronized power measurements<br />Future PMUs+ICTcan provide system-wide synchronized data on ms timescale <br />
  58. 58. Distributed sense & respond<br />SCADA<br />EMS<br />global<br /><ul><li> centralized
  59. 59. state estimation
  60. 60. contingency analysis
  61. 61. OPF dispatch
  62. 62. simulation
  63. 63. human in loop
  64. 64. decentralized
  65. 65. mechanical</li></ul>relay<br />system<br />local<br />slow<br />fast<br />
  66. 66. Distributed sense & respond<br />SCADA<br />EMS<br />Smart<br />Grid<br />global<br /><ul><li> centralized
  67. 67. state estimation
  68. 68. contingency analysis
  69. 69. OPF dispatch
  70. 70. simulation
  71. 71. human in loop
  72. 72. decentralized
  73. 73. software controlled
  74. 74. local alg from global</li></ul> perspective<br /><ul><li>inc. relay system</li></ul>local<br />slow<br />fast<br />What/how to control if info & control<br /> are 1,000x faster?<br />
  75. 75. Where can we help<br />Distributed sense & respond in SG<br />Integration of solar and wind power<br />Pricing and optimization of demand response<br />Architecture and protocols to implement these algorithms<br />These problems are interconnected<br />Each requires engineering+economics<br />
  76. 76. Integration of solar/wind<br />Challenge: solar/wind are unreliable<br />Questions<br />How much renewable power can current grid carry?<br />How can real-time/forecast weather data help predict supply? <br />How (much) can geographical diversity improve reliability?<br />Can EV serve as storage? How?<br />How to do optimal power flow with intermittent sources?<br />
  77. 77. Pricing demand response<br />Questions<br />How to compute optimal prices for demand response?<br />How to price interruptible power for demand response?<br />
  78. 78. Architecture and protocols<br />Questions<br />What protocols are needed to support above control and optimization mechanisms?<br />What architecture features are essential?<br />Where does current architecture fail?<br />Cyber-physical aspect of control and protocols<br />Protocols that enhance cyber-physical security<br />
  79. 79. Outline<br /><ul><li>Snapshot of US power grip
  80. 80. Distribution
  81. 81. Generation
  82. 82. Transmission
  83. 83. Where can we help?
  84. 84. cs/ee 146 projects: open discussion</li></li></ul><li>Research Areas<br />Southern California Edison<br />Systems of systems<br />Software & algorithms for distributed sense and respond<br />Communication systems<br />NSF NetME<br />Theory of network architecture<br />
  85. 85. SCE Info (Mani)<br />Peak demand control<br />Peak hour: 10/year<br />Spread 15 min of peak over 45 min = 1 power plant<br />Voltage control<br />Stable & low voltage (e.g. 105V) at end users = great saving<br />(How) Can real-time control over smart grid achieve it?<br />Workforce: 50% of SCE retiring<br />Automation, machine learning can help<br />Market/demand response: Irvine project<br />What data is available? What does it tell us?<br />
  86. 86. SCE Info (Mani)<br />Battery + solar/wind<br />How much benefit? <br />SCE has a real system and needs answer<br />Protocols<br />Current SCE system: large number of non-standard/insecure/inefficient protocols<br />In addition to ideal protocols, need migration strategy<br />
  87. 87. cs/ee146 Fri class<br />Wed<br />Fri<br /><ul><li> Projects to initiate?
  88. 88. Format and schedule?</li></li></ul><li>cs/ee146 Fri Projects<br />Adam: interested in learning more from SCE about Irvine project and data<br />Andreas: interested in available SCE data on failure for learning and prediction<br />
  89. 89. cs/ee146 Fri class<br />Potential presentations: <br />Solar power: model, literature<br />Simulation: OpenDSS & examples (Julian/Roger)<br />Smart grid protocols, Irvine project (Chris)<br />Electricity market: OPF, transmission rights<br />Net arch/security & smart grid<br />

×