model-based optimization of a solar-powered car · model-based optimization of a solar-powered car....

19
Clean mobility for everyone Arjo van der Ham – Chief Technology Officer May 21 st , 2019 Model-Based Optimization of a Solar-Powered Car

Upload: others

Post on 28-Oct-2019

9 views

Category:

Documents


0 download

TRANSCRIPT

Clean mobility for everyoneArjo van der Ham – Chief Technology Officer

May 21st, 2019

Model-Based Optimization of a Solar-Powered Car

Clean mobility for everyone

Only 3% live near charging points

Leapfrog the grid

Local power generation

1 billion charging points

Coal plant? Upgrade grid

Lightyear One

725kmWLTP range

Solar panels on roof and hood

Solar-electric family car

10.000kmSolar range per year, in NL

60kWhBattery

12km/kWhEfficiency

A two week surf trip in Portugal

0Grams of CO2

1.154 kmTotal distance

6Surfspots

0Charging stops

100+ engineers on a mission200+ years of automotive experience

65Early investors

80+Signed reservations

10Signature editions

€119,000Excluding taxes

The first pioneers

Solar cells are just the start

Energy use vs solar yield

Solar panel efficiency

Solar panel size

Improved aerodynamics

In-wheel motors

Lightweight cycle

Lightweight materials

Re-imagine the car

10

State-of-the-art for SEV:• 3 strings of 125 cells• The weakest cell

determines the string current

Curved Surfaces and Shadows cause mismatches between cells

Therefore, Solar Electric Vehicles have flat solar roofs

This limits design freedom and aerodynamic performance

DC

DC

Grouping Efficiency: ratio between the individual cell MPPs and the

suboptimal group MPPs

11

Solution: Smaller groups, AKA Distributed Maximum Power Point Tracking

𝜂𝑔𝑟𝑜𝑢𝑝 =σ 𝑃

𝑚𝑝𝑝,𝑔𝑟𝑜𝑢𝑝

σ 𝑃𝑚𝑝𝑝,𝑐𝑒𝑙𝑙

𝜂𝑔𝑟𝑜𝑢𝑝 = 75 %𝜂𝑔𝑟𝑜𝑢𝑝 = 98.6 %

PV to Isolated Bus Architecture (Olalla, 2013)

12

Ipv,n

DC

DC

Istr

+

Vbat

-

Ibat

Ipv,2

Ipv,1

+

Vstr

-

DC

DC

In

DC

DC

I2

DC

DC

I1

Ipv,n

DC

DC

Istr

+

Vbat

-

Ibat

Ipv,2

Ipv,1

+

Vstr

-

Ipv,n

DC

DC

Istr

+

Vbat

-

Ibat

Ipv,2

Ipv,1

+

Vstr

-

DC

DC

In

DC

DC

I2

DC

DC

I1

In'

C

Ic

I2'

I1'

12 W

12 W

6 W

30W

2 W

2 W

4 W

10 W

10 W

10 WSteady – State Solution:- All groups at MPP

- Net power to bus is zero

13

MATLAB Solar Simulator Model

CAD Model Solar Array GroupsPlace

cells

K-means

clustering

TMY Weather data

Solar Projection PV to Isolated bus (Steady State)

DC

DC

+V1

-

I1Ipv,1

DC

DC

+V2

-

I2Ipv,2

DC

DC

+Vsec

-

In'Ipv,n

I2'

I1'

IcIn+Vn

-

DC

DC

+Vstr

-

Istr

+Vbat

-

Ibat

𝜂grp 𝛽

Cell IV Curves Group IV Curves

14

Fiat Grande Punto

#cells 129

𝐴𝑠𝑜𝑙𝑎𝑟 1.98 m2

Tesla Model S

#cells 159

𝐴𝑠𝑜𝑙𝑎𝑟 2.44 m2

Stella

#cells 380

𝐴𝑠𝑜𝑙𝑎𝑟 5.84 m2

VW Beetle

#cells 93

𝐴𝑠𝑜𝑙𝑎𝑟 1.43 m2

15

16

Grouping

Efficiency99% 80% 75%

Conversion

Efficiency98% 98% 98%

System

Efficiency97% 79% 74%

Grouping

Efficiency100% 99% 98%

Conversion

Efficiency96% 95% 94%

System

Efficiency96% 94% 92%

DC

DC

+Vbat

-

DC

DC

DC

DC

DC

DC

DC

DC

+Vbat

-

-1 +15 +18

More Curvature

Old

Sys

tem

New

Sys

tem

A look into the future

2019Lightyear One prototypes

2020First deliveries of Lightyear One

20211000 deliveries of Lightyear One

Start Lightyear Two

Key Takeaways

Detailed modelling is key to understanding the design space and creating optimized solutions

For typically curved cars, the DMPPT system can improve performance by up to 18%-points

The diversity of MATLAB / Simulink allows It to be used as a ‘swiss army knife’ for engineers

Clean mobility for everyoneArjo van der Ham – Chief Technology Officer

Thank you