iMinds the conference 2012: Michel Tilman

|

SPARCiMinds 08-11-2012

|

SPARC

Smart Algorithms, security, connectivity

Plug-in EV, PHEV

Automobile Cars (Kangoo, Ampera), scooter (QWIC)

Renewable Charge using wind energy Balance supply and demand

Charging services Platform, algorithms, security, charging stations, mobile

device integration, simulations

8 November 2012 2

|

Participants

8 November 2012 3

REstore

|

Research activities

User profiles Survey and workshops on smart charging and billing of

electric vehicles Clustering of potential EV customers Qualitative analysis of different profiles

Future-proof architecture Scalable platform Pluggable algorithms Steering and monitoring via REST services Mobile device integration Local and remote security services Charging stations supporting real and simulated charging

8 November 2012 4

|

Research activities (continued)

Multi-party security models for identification and billing Triple authentication, authorization, auditing SAML / XACML, secure auditing

Design and evaluation of smart charging algorithms Trivial, Intelligator, Intelligator+, Dual Decomposition Different KPI’s

User satisfaction, balancing supply and demand, green energy usage Simulation framework

Business models and techno-economic studies EV-leasing Smart charging infrastructure

8 November 2012 5

|

Field trial

Setup Parkings Zand and Station (Interparking)

2 charging stations in each parking SPARC servers embedded in master poles

Global server (Sony) Replays real wind data

Test Trivial and smart algorithms Authentication over powerline (Dolphin) Smart connectivity (VITO scooter) Multi-party security Degraded mode operation

Log data

8 November 2012 6

|

Field trial (continued)

78 November 2012

|

Field trial (continued)

8 November 2012 8

|

Field trial (continued)

8 November 2012 9

|

Field trial (continued)

8 November 2012 10

|

ArchitectureZand (Local Server)

RESTful Services

Local Optimizer Scheduler BeCharged Station

BeCharged StationDolphin

Sony (Global Server)

RESTful Services

Global OptimizerWind

TestPlanSimulator / JMeterAndroid App

Dolphin Reader

Kouter (Local Server)

Local Optimizer Scheduler Virtual Station

Virtual StationRESTful Services

QWIC Scooter

iMinds App

Security

Security

Authentication / Authorization / Audit

8 November 2012 11

|

Evaluation KPIs

8 November 2012 12

|

Evaluation KPIs (continued)

8 November 2012 13

|

Evaluation KPIs (continued)Setting• 100 EV’s – 1 month – 3 algorithms

Intelligator: multi-agent system with central coordination. Comfort settings are guaranteed locally.Dual decomposition: local optimization, coordination through virtual prices.

• Realistic behaviour / real wind data

0:00 15:00 6:00 21:0012:00 3:00 18:00 9:00 0:00 14:59 6:00 21:00-50005000

150002500035000450005500065000

Wind Trivial IntelligatorDual decomposition

Time

Pow

er

(W)

KPI1 – User satisfaction

• Are EV’s full (95% or more) at departure?• Yes when departure time is known• Degeneration when leaving early

0% 5% 25% 50% 75%

70%

75%

80%

85%

90%

95%

100%

KPI

Max deviation expected duration

KPI2 – Imbalance reduction

• Works best with diverse profiles• Different algorithms are comparable

KPI3 – Green energy

• Fraction of demand supplied by renewable energy• Algorithms again quite comparable

TrivialIntelligator

8 November 2012 14