Analysis & Optimization of Electricity Infrastructure Hardening Measures

Transcription

1 Analysis & Optimization of Electricity Infrastructure Hardening Measures Workshop for Research in Electricity Infrastructure Hardening Gainesville, FL June 9, 2006 Francis M Lavelle, PhD, PE Applied Research Associates, Inc 8540 Colonnade Center Dr, Suite 307 Raleigh, NC 2765 (99) flavelle@aracom

2 Outline ARA Overview Stochastic Damage Modeling with Example Applications Tree Blowdown Coastal Flooding Research Recommendations 2

3 Applied Research Associates, Inc Founded in 979 Employee Owned,00 Employees FY05 Sales of $49 Million IntraRisk Division Offices 3

4 Wind Engineering Capabilities Over 80 person-years of wind engineering experience Over 200 wind engineering publications Validated wind hazard and vulnerability models for accurate risk assessment Over 20 wind engineering software tools ARA s s Wind Hazard Models Thunderstorms Tornadoes Hurricanes Extratropical Cyclones Extensive post-storm damage investigation experience 4

5 Significant Projects/Activities Transmission line risk studies in US, Canada, & Caribbean Tornado risk assessments for nuclear power plants FEMA hurricane loss evaluation methodology (HAZUS) ASCE 7 design wind speeds for Gulf & Atlantic coasts USACE risk-based levee design criteria development Florida Residential Construction Mitigation Program Building code studies in FL, NC, SC, and TX ASTM wind borne debris standards Worldwide hurricane climatology studies for wind tunnels 5

6 Stochastic Damage Modeling Transmission and distribution systems are highly vulnerable to hurricanes Regional response and recovery efforts are severely impacted by power outages Utilities must find the best balance between costs of hardening, maintenance costs, and outage impacts Historical data and lessons learned are necessary but not sufficient for optimal decisions Stochastic damage models are needed to optimize electricity infrastructure hardening measures 6

7 Hardening Measures Strengthening Pole and tower designs Conductors and conductor connections Maintenance Invest additional resources in inspecting and repairing degraded system components Redundancy Multiple paths 7

8 Optimal Transmission Line Design Decision Analysis Decision Options Design A : MTS-550 (Initial Cost) Design A 2 : MTS-700 Design A n : MTW-700 Design Lifetime Uncertainty States (Tax Credit, Maintenance, and Failure Costs are explicitly associated with each outcome event) P(θ ) Wind Storms P(θ 2 ) P(θ m ) Str Response No Failure Failure Restoration, Outage Costs Monte Carlo Simulation used to estimate event probabilities and compute estimated cost Optimal Design Decision Minimizes the expected value of the total lifetime costs, discounted to present worth * [ ( )] = min EuA ( j ) EuA j =, n [ ] Outcomes _ X(A,θ q ) _ X(A,θ r ) _ X(A,θ t ) Costs u q u r u t t [ ( ) ] = ( θ i ) EuA P u i= i 8

9 SOIL STIFFNESS Modeling Components Event Risk* Wind Fields* Terrain Trees Topography *Hurricane, Thunderstorm, Tornado, Winter storm Line Locations Structures Wind Loads Failure Probabilities Economics Fraction Failed GUY STIFFNESS 4000 Peak Gust Wind Speed Annual Cost ($k) 3000 Design Alternatives 2000 min Annual Failure Probability GROUND LEVEL 9

10 Optimum Risk-Based Design Hurricane Simulation (Windspeed, Direction as Function of Position) 4000 Coastal Region, 40 mile T Line, 5 KV Transmission Line Geometry (Length, Orientation, Geographic Location) Wind Load Estimates and Failure Probabilities for Design Alternatives Expected Total Annual Cost ($k) min Steel A Steel C Steel B Wood A Steel D Steel E Steel C Wood C Steel F Wood D Wood B Curve of Minimum Total Cost for Selected Designs Aggregate all Costs and Discount to Present Worth Annual Failure Probability 0

11 Minimize Total Lifetime Cost Construction Cost Expected Total Annual Cost ($k) Construction, Tax Credit, and Maintenance Cost Total Cost Failure Cost Sensitivity Analysis on Lost Revenues MTS-550 MTS-700 MTS-800 Annual Failure Probability MTW-700 Design Decision Options

12 Decision Threshold Analysis on Failure Cost Sensitivity MTS Steel 550 MTS Steel 700 MTS Steel 800 Expected Expected Value Value Lifetime Lifetime Costs Costs ($000) ($000) - - PW PW $/kw-hr $/kw-hr Load Load (MW) (MW) 2

13 Tree Blowdown Load and resistance model based on research performed by the USFS in the 950 s and additional work performed by ARA in the 980 s Models probability of uprooting or stem failure as a function of Tree characteristics and tree density Peak gust wind speed Tree inventory database developed for contiguous US and implemented in HAZUS-MH Hurricane Model for Tree damage to property Tree debris removal estimation 3

14 Tree Load and Response Model Tree is modeled as a SDOF system: me && x + cx &&+ kx = FD ( t ) Stiffness: K 3EI = H T = a bh 3 bh + b Kψ H d 2 bh bh ( c, f ) Period of vibration: +ε s Ψ FD ρu 2 D = k 2 k = 2 k d 2 + k 3 c h 2 W R W dbf dc R W dc Ψ D 5 5 where the drag parameters k and k2 are modeled as lognormal random variables that depend on the modulus of rupture 4

15 Example Simulation Results Base Bending Moments (kn-m) Mean RMS Peak 5

16 Model Flow in Tree Canopies 25 CdLAI=03, Zo/H=04, d/h= 2 Non-Dimensional Height (z/h) 5 0 Turbuelnce Intensity ESDU Mean Velocity Mean Velocity (U(z)/U(H)) or Intensity 6

17 Simulation Approach Sample tree parameters 2 Choose value of C d LAI for forest 3 Compute effective velocity parameters 4 Generate time series of wind speeds (0 minutes) 5 Compute value of C d A for sampled tree 6 Compute minimum failure wind speed and convert to open terrain equivalent 7 Repeat 00 times 8 Compute effective number of stems/hectare γ = 4 0 C C d LAI A d 7

18 Example Probability of Blowdown 70 Foot Deciduous Tree Probability of Blowdown CdLAI=030 CdLAI=020 CdLAI= CdLAI=00 CdLAI=005 CdLAI= Peak Gust Wind Speed (mph) 8

19 Example Probability of Blowdown 70 Foot Coniferous Tree Probability of Blowdown CdLAI=030 CdLAI=020 CdLAI= CdLAI=00 CdLAI=005 CdLAI= Peak Gust Wind Speed (mph) 9

20 Validation of Tree Blowdown Data collected following Hurricane Isabel (2003) Eight randomly selected areas in Northeastern North Carolina Number of trees on each lot By height class By tree type (deciduous or coniferous) By performance (uproot, stem failure, no failure) 20

21 Summary of Data Collected Location Peak Gust Wind Speed (mph) Number of Lots Surveyed Total Number of Trees % of Trees Blown Down Ahoski % Ahoski % Elizabeth City % Elizabeth City % Manteo % Manteo % South Mills % Windsor % Total % 2

22 Deciduous Trees Probability of Blowdow CdLAI=005 CdLAI=00 Isabel Tree Data 70 Foot Deciduous Tree Probability of Blowdow CdLAI=005 CdLAI=00 Isabel Data 50 Foot Deciduous Tree Peak Gust Wind Speed (mph) Peak Gust Wind Speed (mph) 35 Foot Deciduous Tree Probability of Blowdow CdLAI=005 CdLAI=00 Isabel Data Peak Gust Wind Speed (mph) 22

23 Coniferous Trees 70 Foot Coniferous Tree -- Shifted 0 mph 50 Foot Coniferous Tree -- Shifted 5 mph Probability of Blowdown CdLAI=005 CdLAI=00 Isabel Data Probability of Blowdown CdLAI=005 CdLAI=00 Isabel Data Peak Gust Wind Speed (mph) Peak Gust Wind Speed (mph) Probability of Blowdown Foot Coniferous Tree -- Shifted 30 mph CdLAI=005 CdLAI=00 Isabel Data Peak Gust Wind Speed (mph) 23

24 Coastal Flooding Risk Analysis Model components Storm surge Astronomical tide Wave set-up and run-up Applications Planning, design, & mitigation Emergency response Insurance Latitude /09/6@07: Longitude Hurricane Ivan storm surge and waves in Escambia Bay

25 Coastal Flooding Model Overview Storm Track, ΔP, Rmax, B HURRICANE WIND FIELD MODEL Ten-meter Winds DEEP-WATER WAVE MODEL MODEL (WAVEWATCH III) STORM SURGE/TIDE MODEL (Finite Difference) SWAN Wave Model SETUP/BREAKING SURGE/TIDE/SETUP Total Water Level Ten-meter Winds DAMAGE POTENTIAL TO COASTAL STRUCTURE Wave Breaking/Runup 25

26 Hurricane Katrina Storm Surge 26

27 Katrina Storm Surge Prediction 27

28 Katrina Wave Height Prediction 8 6 Data SWAN Buoy 42040, Hurricane Katrina 4 2 Wave Height (m) :00 6:00 2:00 8:00 0:00 6:00 2:00 8:00 0:00 Time UTC, 8/28/2005-8/30/

29 Research Recommendations Transmission line optimization studies Optimization of new construction and maintenance Impacts of response and recovery costs and indirect economic losses on design and maintenance decisions Distribution line optimization studies Regional analysis via statistical analysis at block level Analysis of above ground vs below ground installation Develop and validate models for tree damage to distribution lines Coastal flooding impacts on power plants, substations, and T&D systems 29