feasibility studies for the introduction of electric vehicles in - - PowerPoint PPT Presentation

Using mobility information to perform feasibility studies for the introduction of electric vehicles in taxi fleets

Jesús Fraile Ardanuy

ETSI de Telecomunicación Universidad Politécnica de Madrid Hasselt, July 13th 2015

Who am I?

Jesús Fraile-Ardanuy

Associate Professor Technical University of Madrid

Outline

• Introduction
• Fundamentals of Electric Vehicles (EVs)
• Big Data and EVs
• Data Mobility description
• Results
• Other lines of work
• Conclusions

INTR NTROD ODUCTIO UCTION

World Population

World population forecast

• Rates of populatio

ation growth th are currently highest in the less developed regions.

• If curre

rent nt trends ds contin tinue ue:

– Africa’s share will rise to 20% – Asia’s population will decrease slightly to 57% of the world total in 2050. – Europe’s share will drop below Latin America’s.

http://www.theguardian.com/world/2011/jan/14/population-explosion-seven-billion

Urban and rural population

• Globally, more people live

in in urban areas than in rural al areas.

– In 2007, the global urban population exceed the global rural one.

• Level of urbanization varies greatly across regions.

http://esa.un.org/unpd/wup/Highlights/WUP2014-Highlights.pdf

Urban-rural population

http://image.guardian.co.uk/sys-files/Guardian/documents/2007/06/27/URBAN_WORLD_2806.pdf

• Africa and Asia are urbanizing more rapidly than other

regions in the world.

Urban and rural population

http://www.populationlabs.com/world_population.asp

Urban population problems and challenges

• There are many problems associated with the rapid urban population

growth:

– Unplanned housing

http://www.geo.tv/article-112485-Traffic-jam-in-Karachi-residents-forced-to-open-fast-on-roads- http://urbanpoverty.intellecap.com/?p=552 http://www.china-mike.com/facts-about-china/facts-pollution-environment-energy/

– Water and energy – Urban waste – Stress on the infraestructure – Basic services: education and health care. – Pollution

Air pollution in Urban areas

http://aqicn.org/map/world/

Problems

• The biggest threat to clean air these

days is traffic emissions.

• Cars are responsible for 73% of

urban air pollutants.

• Petrol and diesel-engined vehicles

emit a variety of pollutants, principally carbon monoxide (CO),

• xides of nitrogen (NOx), volatile
• rganic compounds (VOCs) and

particulate matter (PM10).

• The pollutants have linked to

chronic health problems like asthma, lung cancer, emphysema, and heart disease.

http://uk-air.defra.gov.uk/air-pollution/effects http://newhamgreenparty.com/2015/03/15/tackling-air-pollution/ http://newhamgreenparty.com/2015/03/15/tackling-air-pollution/

The reducti ction n of polluta tant nt emissi ssions ns and improving

• ving air quality

ty in urban areas s are fundamen amenta tal l aspect cts s to be solved in the followin wing g years

Solution lution?

Strategies to EV deployment in urban areas

• Strateg

rategies es:

– Promoting public transportation, bicycling and walking in the cities, reducing the number of vehicles in the streets.

http://studyinuk.universiablogs.net/2013/11/05/take-a-walk-to-the-campus/image0017-584x234/ http://bellovelo.blogspot.com.es/2010/02/great-bike-friendly-cities.html

– Promoting the transition from ICE  EVs.

http://www.wired.com/tag/tesla-model-s/

EV promotion

• Governments have been promoting EVs

through different initiatives:

https://en.wikipedia.org/wiki/Government_incentives_for_plug-in_electric_vehicles http://www.greenwisebusiness.co.uk/news/transport-for- london-issues-67m-tender-for-green-vehicles- 1320.aspx#.VZPvHfntlBc http://www.plugincars.com/public-charging-why-its-time-think-plugging-127217.html

– Subsides to purchase EVs. – Creation of ultra-low emissions zones (ULEZ) in city centers. – High occupancy vehicle (HOV) lane access – Tax exemptions and other fiscal incentives – Priority parking – Insurance discounts – Deployment of charging points

EL ELEC ECTRI TRIC C VE VEHI HICL CLES ES

But…What is an electric vehicle?

• An EV is a vehicle that uses one (or more) electric motors

for propulsion, instead of an ICE.

http://treneando.com/2012/01/23/parla-apuesta-por- el-tranvia-que-en-2011-supero-los-cinco-millones-de- usuarios/ https://movimientoindignadosspanishrevolution.wor dpress.com/el-ave-no-es-rentable-en-espana/ http://www.motoryracing.com/camiones/noticias/scania-siemens- trabajan-camion-electrico-perfecto/

• An EV can be powered:

– Through a collector system by electricity from off-vehicle – Self-contained using a battery or generator to convert fuel to electricity.

Electric vehicle classification

Motor/ Generator

Battery Fuel

Transmission

Engine

Fuel Transmission Engine Battery Transmission

Motor/ Generator

Battery Electric Hybrid Conventional

Electric Vehicles 101. Dan Lauber MIT

Hybrid EV

• Hybrid EV

– ICE+electric motor-generator

• Fueled by gasoline, diesel,

compressed natural gas or bio-fuels – Small battery pack – Recharged from regenerative braking – Limited all-electric range (2-3 km) – No support pport exter ternal nal charging arging (no no plug-in) in) – ICE engine gine mor

• re

e powe werful ul than EV mo motor

• r

– Types:

• Micro

ro hybrid id (stop & start)

• Mild hybri

rid (assisst to ICE)

• Full hybrid

id (Electric motor can drive the car)

http://www.taringa.net/posts/autos-motos/18323659/Toyota-Prius.html http://www.lexusofglendale.com/los-angeles-hybrid-lexus

Plug in Hybrid Electric Vehicle

• Plug in Hybrid EV

– ICE+electric motor-generator – Larger battery pack – Recharged from regenerative braking and external charging – Limited all-electric range (25-50 km) – ICE engine>EV motor

http://www.autoblog.com/2014/04/28/2015-bmw-i8-review-first-drive-video/ http://cocheselectricos365.com/mitsubishi-outlander-phev-plug-electric-vehicle-13247.html http://www.hibridosyelectricos.com/articulo/mercado/nuevo-bmw-x5-xdrive-40e-hibrido- enchufable/20150316133159009036.html

Plug in Hybrid Electric Vehicle

• Ex

Exten ended ed-Pl Plug ug in in Hybrid EV

– A PHEV with bigger battery – Driving ranges (60-100 km) – All electric mode in day-by-day activities – ICE engine<EV motor – ICE engine is added to an EV motor to charge battery (no for propulsion)

http://www.opel.es/vehiculos/coches-opel/vehiculos-de-pasajeros-opel/ampera/models/available-models.html

Degrees of hybridization

Source: http://www.hybridcenter.org/hybrid-center-how-hybrid-cars-work-under-the-hood.html

Efficiency

Micro Hybrid

Citroën C2

Mild Hybrid

Honda Insight

Full Hybrid

Toyota Prius

Ext- PHEV

Chevy Volt

Plug-in Hybrid

BMW i8

Pure Electric Vehicles

• BEV:

: Batte tery ry Electric Vehicle

– No ICE – Different ranges depending on nominal battery capacity:

• iMiev (150 km)
• Leaf (170 km)
• Model S (350 km)
• E6 (400 km)

– No plan B if you are out of battery!

Benefits of EVs

• More efficient.
• Lower energy cost compared

to oil.

• Lower emissions (depending
• n the country)

– But it is easier to control emissions at few large locations (power plants) than millions of tailpipes

• Simpler transmissions. Fewer

moving parts.

• Noise reduction.

Current Challenges

• Limited range

– Large battery (weight/size)

• Long charge times
• High initial cost
• Battery life
• Consumer acceptance
• Grid Integration

Consumer acceptance

http://www.continental-corporation.com/www/download/pressportal_com_en/themes/initiatives/channel_mobility_study_en/ov_mobility_study2015_en/download_channel/mobistud2015_praesentation_en.pdf

Are EV boring?

Go to min 4

Understanding power systems

Thermal power plant Hydro power plant Wind Energy Transmission Substation System Operator (SO) control center Transmission NETWORK Distribution substation Residential Customers (Low Voltage) Industrial Customers (Medium or High Voltage)

Generation

Energy flows in one direction, from generation to consumer at the lowest cost and at the highest reliability

Source: REE.es

Balance between generation and demand

30/08/2015

BI BIG DATA G DATA AND AND EVS EVS

Big Data and EVs

• Big data applied to EVs can turn the information from

the vehicles in into meaningful ingful operationa ational in insig ights hts and insights hts about about the customer’s behavior

• r.

Improving human experience

 Personal level observation

Charging the EV has a significant cost since it was done during peak load period. Consider changing this time to night period.

EV monitoring application Action in the physical world

 Population level observation

Source of Big Data in Power Systems

From BD in the management of EES. Louis Wehenkel

• Observat

ervationa nal datasets asets

– Meteo

• Wind, rain, clouding, temperature, etc.
• Measurable at any place and at any time
• Influences demands, offers, harzards, equipment ageing
• Simulated

ated datasets asets

– Generated and used to replace or forecast unavailable observational quantities – Economics

• Prices, bids, costs of consumers and producers
• Measurable for any actor and at any time
• Influence system technical and economic performance

– Technical performance

• Failures, power flows, service disruptions, quality

Big data in EVs

Sources of Big Data in EVs

• Cars are generating lots of data every second:

– Acceleration/Braking/g forces – Idle / Number of stops – Electric Consumption/Battery state – Ambient temperature – HVAC temperature – Tire preassures – GPS traces – Charging time periods – Charging rated power

How to use these data?

Electric Vehicle DATA

Drivers Fleet managers Retailers DSOs TSO Generators

What Big Data can do for drivers?

• Improving driving effi

fici cien ency cy

• Allowing to detect

ct anomali alies and problems in their own vehicles

• Optimizing charging electr

trici icity ty costs sts

Improving driver behavior. Big Data for drivers

• Provi

viding ing feedb dback ack to drivers ers on how they are currently doing.

• Comparing to

personal historical data

• Comparing to other

similar drivers

– Same mobility patterns (same route

• r area)

– Same type of EVs

Improving driver behavior. Big Data for drivers

• Providing feedback to drivers on how

how to do it it bett tter er.

– Avoiding aggresive aceleration/braking events – Time spent idling

Detecting problems in the Vehicles. Big Data for drivers

• Sharing and comparing different measurements will

allow to id identify ify anomalo alous us vehic icle le behavior vior.

– Anor

• rma

mal range ange reduct eduction ion or higher gher averag erage battery ery tempera perature ure can lead to battery problems (accelerated ageing

• f the battery)

– Diagn gnos

• stics

ics trouble uble codes

• des

Optimizing charging electricity costs. Big Data for driver

• Optimizing charging costs, taking into account:

– Vari riab able le electric ctricity ity pric ice – The persona rsonal daily ly sche hedule ule

• Determining posib

sible le charging arging locati ation

• ns
• Determining posib

sible le charging arging period riods

What can Big Data do for fleets?

• Comprenhensive analysis of:

– Fuel el/elec electrici tricity ty econo conomy repor eporting ing

• Measuring real-world consumption

from all fleet vehicles. – Idle monit nitori

• ring

ng and d mana nagem gement ent

• Reporting idle periods and allowing to

quantify savings and identify drivers that may require additional route adjustments

What can Big Data do for fleets?

• Driver behavi

vior

• r feedb

dback ack

• Di

Diagnostic stic troubl

• uble

e codes

• Distribution of charging

ging times es

• Ve

Vehic icle le lo locati tion n tracking

What can Big Data do for Elec. Retailers?

• EV Load fore

recast casting

– Improving their offer bids and increasing their benefits

• Segmentation

entation-dri driven ven marketin keting g offers

• Special

al tariff iff designs

What can Big Data do for DSO?

• More

e effe fecti ctive ve monitor toring ng and proactive maintenance

– Obtaining operation conditions for charging EVs on local household distribution grid.

• Power losses
• Power quality (voltage and current profiles, unbalance and

harmonics) – Modell delling ing large ge scale ale (spatial-temporal) deployment of EVs and quantifying the impacts on distribution operation conditions and infrastructures.

What Bid Data can do for DSO?

• Investigating optimal EV

charging profiles that result in maximal economic, environmental benefits and minimal operation disturbance.

• Reducing or postponing

the need for network reinforcement through charging active demand management.

What can Big Data do for TSO?

• Oper

erati ation

• n (sh

short

• rt term

rm)

– Predict ict network

• rk flow over the next minute

utes, s, hours, , days and weeks. – Optimize imize the power system tem accordingly (tradeoff between reliability-economy)

From BD in the management of EES. Louis Wehenkel

• Asset

et mana nagem gement ent (mid id term)

– Understandi rstanding facto ctors rs driving aging and failures lures of components – Undestan stand critically tically of componen

• nents

ts’ availability lability for system operation – Optimi imize ze the repairing ring and replacemen acement of equipment accordingly

• Inves

estm tment ent (long

• ng term)

– Predict ict usage of the power system tem over the next years – Accordingly, take highly strate rategi gic importan rtant decision sions

What can Big Data do for Generators?

• EV Load fore

recast casting

– Improving their supply bids and increasing their benefits

• Integration of intermittent generation
• Combined generation bids:

– Wind energy + distributed storage capacity of EVs

EV EV A AND TAX ND TAXI FL FLEE EETS TS

Benefits of EV in taxi fleet

• Improvements in Air Quality

– EVs have zero tailpipe emissions. – Highest GHG concentrations is found in areas with high traffic rate (also high density of taxi trips). – 20% of total generated electric energy in California comes from renewables. – Using cleaner energy sources will reduce the emissions asssociated with powering EVs.

Benefits of EV in taxi fleet

• Reduced Carbon footprint

– Even after accounting for the energy- production-level emissions associated with Evs, electrification of taxis would lower the fleet’s carbon emissions.

• Resiliency

– EV can also be designed to be usable as mobile power storage units in the event of an emergency

Benefits of EV in taxi fleet

• Visibility
• Price consistency

– Electricity prices are much less volatile.

• Energy security

– Reducing the country petroleum imports.

Economics of EV ownership

• Factors that determine the adoption of an EV:

– Vehicle icle pur urchase chase price ice

• Battery price is the key driver of purchase price
• Price is projected to decline over the time

– Maint intena enance nce and d repa pairs irs

• Lower maintenance (savings)

– Batter ery replacem placement ent

• Battery is degraded over time, more quickly if is quick charged

(need to be replaced)

• Oppor

portunity tunity: battery reuse for static applications. – Year ars vehic hicle le in ser ervice vice – Resid sidua ual value lue – Cost st of elec ectric tricity ty

• Taxi operators, to maximize

revenues and minimize downtime, limiting time available for charging.

DA DATA TA INFO NFORMA RMATI TION ON

Data information

• General information:

– GPS traces 466 Vehicles of Yellow Taxi Cap. – Collected: May-June 2008. – Data provides:

• Lat-Lon
• Time stamp (Unix time)
• Ocupation

Data information

• It is assumed:

– We are focus on

• n the taxi

xi (no on the driver).

• Taxis can be driven

en by by differ ferent nt drivers ers. – Drivers have same skills (similar knowledge of the city).

Consumption model

• EV consumption model

Input ut: GPS trac ack Consider:

• Terrain elevat

evation ion

• Auxi

xili liary ary loads ds (lights/heating)

• Occu

cupat ation ion (increasing mass for vehicle occupied with customer) Output put:

• Power
• Cons

nsumed med energy ergy

• SoC evolut
• lution

ion

Consumption model

Consumption model

 

 cos g M M R F

d car rr

 

2

2 1 v AC F

d a

 

 

 sin g M M F

d car hc

 

 a

M M F

d car la

  05 . 1

la hc a rr te

F F F F F    

• Equations:

v F P

te te 

Consumption model

• Forward driving:

gear te

• ut

mot

P P  

_

mot

• ut

mot in mot

P P 

_ _



aux in mot bat

P P P  

_

v F P

te te 

Consumption model

• Regenerative braking:

te ratio gen reg te

P R P

_ _



reg te gear

• ut

mot

P P

_ _

 

reg te gear

• ut

mot

P P

_ _

 

aux in mot bat

P P P  

_

Consumption model

• Battery dynamics:

– Discharging process (moving forward) – Charging process (regenerative braking)

Consumption model

• Results

2000 4000 6000 8000 10000 12000 14000 16000 18000 20 40 60 80 100 Time [s] 2000 4000 6000 8000 10000 12000 14000 16000 18000 20 40 60 80 100 State of Charge (%) speed (kph) Car Stopped (Speed=0 kph & SoC=58%)

Battey SoC (%)

Consumption model validation

• We have tuned our model based on real test consumption:

– 111.4 km (69.2 mile) – 3.9 miles/kWh – 0.159 kWh/km

http://insideevs.com/real-world-test-2013-nissan-leaf-range-vs-2012-nissan-leaf-range/

Our consumption model: 0.165 kWh/km

Consumption model

• More complex Consumption models are available

http://vbn.aau.dk/files/55733132/Electric_Vehicles_Modelling_and_Simulations.pdf

Transmission Electric Machine Inverter (Power Electronics) Battery

Consumption model

http://vbn.aau.dk/files/55733132/Electric_Vehicles_Modelling_and_Simulations.pdf

RE RESU SULT LTS

Results

• Analyzing the spatio-temporal mobility of a single taxi vehicle.

Results

• Vehicle: 1
• Number of days: 24
• Number of movements: 49
• Number of stops

ps (>30 min): 48

> 30 min

Results

>30 min

Empty Occupied

Results

Results

• How is the distribution of distance travelled

between two consecutive stops (> 30 min)?

Results

• Electrification Rate: 63.27%

Results

• St

Stop locati ation

• n and durati

ation:

Results

• Best location for charging points

Results

• How long are they stopped? Histogram stop

p time e durati ation

Results

• St

Stop locati ation

• n and St

Stop Initi itial al Ti Time:

Results

• When are they parked? Histogram stop
• p initia

tial time

Results

• How much energy are they demanding?

Results

• Energy demanded during the recharging process:

297.21 kWh

Results ELECT CTRIC RIC TAXI CONVENTIONAL ENTIONAL TAX AXI

Results

• Total energy demanded: 297.211 kWh
• Electricity Price: 23.3 cents/kWh
• Total distance: 1,325.5 miles
• Total distance: 2,132.8 km
• Total cost: $69.25

9.25

• Gasoline Price: $3.692/gallon
• Consumption: 16 miles/gallon
• Total distance: 1,325.5 miles
• Total distance: 2,132.8 km
• Total cost: $244

44.56 56

• Saving: $175.31

http://www.bls.gov/regions/west/news-release/averageenergyprices_sanfrancisco.htma

ELECT CTRIC RIC TAXI CONVEN ENTIONA TIONAL TAXI

Results

• Centroid: (Lat, Lon):

37°46'42.2"N 122°25'01.6"W

• Radius of gyration:

2.56 km (1.6 mil)

Pick up points

Results

• Centroid: (Lat, Lon):

37 37°46 46'27. '27.5"N 5"N 122 122°24'59 4'59.8" .8"W

• Radius of gyration:

3.18 km (1.97 mil)

Drop off points

Results

• When are taxis occupied?

Results

• Average speed:

16.13 km/h Empty taxi

Results

• Average speed:

26.92 km/h Occupied taxi

Results

• Average distance:

3.52 km Empty taxi

• Average distance:

4.2 km Occupied taxi

Results

Results

• Analyzing the spatio-temporal mobility of a taxi fleet.

Results

• Number of analyzed Vehicles: 466 Taxis.
• Average number of days analyzed: 23 days
• Average number of stops > 30 min: 60

Results

• Gyration radius distribution for empty and
• ccupied situation.

km 785 . 91 . 3  

• ccupied

gyr

r km 4 . 1 51 . 3  

empty gyr

r

Results

• Time duration of the stops.

– Max: 17 days (413.5 hours) – Stop (>30 min) less than 24 hours: 99.36%

• Average time duration: 2 hours 34 minutes.

Results

• Starting time to recharge EV taxis:

Results

• Energy demanded by all vehicles:

Results

• Energy recharged during the stops: 170

170 MWh

• Electrificability rate: 65.3% of the total journeys

Battery Capacity: 24 kWh

Results

• California daily electricity demand

http://www.caiso.com/outlook/SystemStatus.html

Results

• Impact on the California daily electricity demand:

0.002% in the peak (14:00)

AD ADDI DITI TIONA ONAL RE RESE SEAR ARCH CH

• A mobile application for identifying the potential for EV

adoption in company fleets.

• The app records:

– Distance traveled. – Average speed. – It is posible to record energy consumption.

New developments at ETSIT-UPM

• The app provides the user with

information about the daily driving distances can be cover using an electric vehicle.

• A database with the technical

specifications of different EVs are used to advice users.

Mobile app description

• Initia

tial screen een:

1. Start to register 2. Analyzing a track 3. Analyzing all tracks 4. Share the track

1 1 2 3 4 4 3 2

Mobile app description

• From the recorded data (single track), the app provide

to the user with the following information:

– Electrificability (yes/no) – Speed – Electricity cost – Fuel cost – Number of subtracks – Total distance travelled – Total saving

Mobile app description

• From the recorded data (all tracks), the app provide to

the user with the following information:

– Electrificability (percentage) – Number of tracks – Electricity cost – Fuel cost – Average distance per track – Total distance travelled – Total saving

Mobile app description

• Configuration screen:

Improvements

• Provi

vidin ing feedb dback ack to drivers ers on how they are currently doing.

• Comparing to personal

historical data

• Comparing to other similar

drivers

– Same mobility patterns (same route or area) – Same type of EVs

CO CONCL NCLUS USIONS IONS

The future

• New applications can modify traditional business.

The future

• Recent Paper:

– Greenblatt, J. B., Saxena, S. “Autonomou nomous taxis could d greatl tly reduc uce e greenhou nhouse-gas gas emissions

• ns of US light

ht-duty duty vehicles es”, Nature Clim. Change, 2015/07/06/online http://dx.doi.org/10.1038/nclimate2685

• Self-driving 'taxibots' could replac

place 90% % of car ars

• Drive

iverless less cabs abs will dramatically ease se congest ngestion ion in major jor cities es

• Even with only

ly one e passen ssenger ger per r ride, car number dro ropped pped by 77 %

• Swapping personal cars with self-driving cabs would free valuable

space.

The future

Questions?