Pecan Street Inc SWEDE 2013 May 5 2013

Pecan Street, Inc. SWEDE 2013 May 5, 2013 Austin, Texas Center for Electromechanics P ECAN S TREET P ROJECT – A “G REEN” D EVELOPMENT Dr. Fabian Uriarte Center for Electromechanics University of Texas at Austin

S MART G RID L OCATION Pecan Street offices Ctr. for Electromechanics Texas Advanced Computing Center pecanstreet. org Mueller Community

Homes 735 Transformers 94 Solar panels 200 Electric vehicles 100

T RANSFORMER L OAD* Circuit feed (7. 2 k. V) 20 -40 xfms. Homes per transformer: Maximum: 11 Mode: 8 Minimum: 4 240/120 V *One possible combination

L OAD D ISTRIBUTIONBY P HASE 333 216 186 350 300 Count 250 60 200 150 21 32 Homes 54 33 100 50 86 (Photovoltaic Arrays) 19 45 28 EVs 0 Phase A PVs (Electric Vehicles) Phase B Phase C Xfms

C ASE S TUDY Center for Electromechanics 800 735 700 600 Simulation type Phasor Stop time 1 day Time step 1 min Num. of time steps 1, 536 Run time 10 s 500 400 300 178 200 106 100 0 Homes * Modeling in progress PVs EVs 98 Cables 94 Xfms 50 1 Residential Community Storage*

S IMULATION A PPROACH Use Recorded Data as Input Run Model Grow Examine Results

R ESIDENTIAL L OAD Real Data

PV G ENERATION Real Data 0. 5 - 1 MW of distributed PV generation daily

EV L OAD Mix of 120 V and 240 V charging > 4 PM

T RANSFORMERS T HROUGHPUT

T RANSFORMER U TILIZATION Before PVs and EVs After PVs and EVs • High PV/Load ratio causes reverse flows • Can increase or decrease xfm. utilization

C HANGE IN T RANSFORMER L OAD No change Example (blue areas): • Usage before: 25 % (fwd. direction) • Usage after: 5 % (reverse direction) • Change: 20 % • “-” or “<0” shows reduction in utilization • “>0” means increase in utilization PVs produce 23% increase EVs produce 2% increase

As seen from here L ATERAL P OWER Before PVs and EVs After PVs and EVs PF PF VA VA Watts Vars • 0. 5 MW reduction in real power demand • Equal reactive power demand • Power factor drops

L ATERAL C URRENT From here Before PVs and EVs After PVs and EVs • Decrease in diurnal current unbalance • Peak time unbalance unchanged

D ISTRIBUTION L OSSES Before PVs and EVs After PVs and EVs Increase due to EVs Reduction due to PVs Xfm losses almost unchanged $45/day = $16 k/year

Center for Electromechanics Simulation Model C ONCLUSIONS factor We have real data (1 m resolution ) Consumption PV generation Can simulate entire smart grid (735 homes) Confidence in results Transformers Appear oversized already Some operate ~80% Can meet EV load (at Mueller community) Power flow is forwards and backwards Residential Solar Panels (PVs) Inject power back into the grid Injection is uncontrolled and unbalanced Reduce lateral and transformer power Provide voltage support Electric Vehicles (Chevy Volts) Uncontrolled charging exacerbates peak demand Electrical impact appears small due to transformer sizing

A CKNOWLEDGEMENTS Pecan Street, Inc. UT Austin Energy • Brewster Mc. Cracken executive director • Dr. Kwasinski & Amir Toliyat solar arrays • Kurt Stogdill utility strategist • Bert Haskell technology director • Ariane Beck program manager • Chris Holcomb data analysis • Dan Weisberg economic strategy • Dr. Baldick & David Tuttle electric vehicles • Dr. Edgar, Wesley Cole, Robert Fares, and Akshay Sriprasad energy storage • Charles Robinson distribution planning • Sharon Bickford systems engineering Center for Electromechanics • Robert Hebner director • John Herbst program manager

Q UESTIONS Center for Electromechanics Fabian Uriarte f. uriarte@cem. utexas. edu Center for Electromechanics The University of Texas at Austin utexas. edu/research/cem 10100 Burnet Road, Austin, TX