You are on page 1of 25

SolarPower Webinar

PV Cleaning: Choosing the


Optimal Method and Frequency

18 October 2017

Partner
PV Cleaning: Choosing the Optimal Method and Frequency

MODERATOR
Michael Schmela

Executive Advisor

SPEAKER
James Kurz

Project Manager

Partner
PV Cleaning: Choosing the Optimal Method and Frequency

Solar Power Europe Webinar – October 18, 2017


James Kurz, Project Manager
Huge PV power plants are being built in desert
regions around the world at highly competitive
generation cost (increasingly < 3 cents/kWh)…

Source: Google Maps (Dubai DEWA Sheikh Maktoum Phase 1 & 2 power plants)

2
…but there is a very dirty elephant in the room:
PV module soiling.
3
3
Cleaning is the only solution in many regions.

4
Urgent question for IPPs: what is the
commercially optimal, least-risk way to deal
with soiling?

5
Apricum – The Cleantech Advisory.
Apricum at a glance
Business Founded in 2008, over 200 successful
transaction advisory and strategy consulting
projects

Industry focus Cleantech. Strong focus on solar, wind,


water, energy storage and digital energy

Team >40 cleantech experts with decades of


industry experience

Clients Companies, investors and public institutions

Services Transaction advisory


Strategy consulting

Locations HQ in Berlin, Germany


Branch offices: Abu Dhabi and Dubai
Representative offices: Argentina, Brazil,
Mexico, USA, UK, The Netherlands, Turkey,
Saudi Arabia, India, China, Thailand,
Indonesia, Philippines, South Korea, Japan

6
ILLUSTRATIVE

Cleaning is necessary in high and certain moderate soiling


regions.
Soiling Temperature [°C], Share of global Need for cleaning
category precipitation [mm] installations1
High soiling Constantly arid, Regular and/or frequent
precipitation ~0
50 °C mm 100
25 50
0 0
Jan Dec

Moderate Pronounced arid Infrequent; often only in dry


soiling 50
season
100
season
25 50
0 0
Jan Dec

Low soiling Relatively humid


Cleaning rarely required
throughout the year unless particular
50 100 circumstances persist (nearby
25 50
heavy agriculture, bird
0 0
Jan Dec droppings, pollution, etc.)
Source: Apricum analysis; 1) share of cumulative ’17–’20 PV market

7
PV projects are nowadays considered highly secure
investments with relatively low equity returns.
Factors affecting the equity return of PV projects

Financial
Project revenue leverage
• PV yield (x3–6)
• Tariff
Equity IRR
(6–9%)1
Project cost
• Cost of construction
• O&M cost

1) With highly creditworthy counterparties, e.g., in the UAE

8
PV module soiling and cleaning affects project returns
substantially through multiple drivers.
Factors affecting the equity return of PV projects

Soiling reduces
PV yield, partly
mitigated by
Financial
cleaning
Project revenue leverage
• PV yield (x3–6)
• Tariff
Equity IRR
Cleaning requires (???)
upfront Project cost
infrastructure • Cost of construction
• O&M cost
Cleaning incurs
running costs,
escalating over
time

9
Developers take substantial risks, underestimating the
uncertainty around soiling.
Factors affecting the equity return of PV projects

Soiling reduces
PV yield, partly
mitigated by
Financial
cleaning
Project revenue leverage
• PV yield (x3–6)
• Tariff

Cleaning requires Equity IRR


upfront Project cost
infrastructure • Cost of construction
• O&M cost Residual uncertainty in yield
and cost impact, highly
Cleaning incurs leveraged, creates
running costs, potentially large impact on
escalating over equity IRR. Little safety
time margin left!

10
Cleaning strategy is a major element in highly competitive
PV project tenders in desert locations.

Developer wants to bid On top of stochastic mean of


lowest tariff possible to win necessary tariff, developer
the tender wants as little safety margin
as possible:

Cleaning strategy needs to be:

Return-optimizing Robust
• Optimal cleaning method • Minimum residual yield
• Optimum cleaning regime uncertainty
(frequency) • Minimum cost uncertainty

11
First and crucial step is to understand climatic and
economic parameters of the site.
Typical site-specific parameters

Climatic Economic Site geometry

• Soiling rate under regular • Labor cost • Site area


conditions • Inflation rate • Site dimensions
 Average daily soiling rate • Water cost • Number of access roads
 Asymptotic soiling rate • Cost for water infrastructure through plant in each
• Sandstorms (e.g., large water tanks) orientation
 Distribution and intensity • Diesel cost (for truck-based)
• Rain events
 Length of the wet/dry
seasons
 Average number of rain
events in wet/dry
seasons

12
Next the plant layout and mounting structure architecture
need to be accounted for.
Layout specific parameters

Mounting architecture Module orientation Other

• Fixed tilt • Portrait • Possibility to connect tables


to form long rows
• Ability to stow tracker
systems at high angles
• Tracker with distributed • Landscape
drive system • Gaps between tables in a
row
• Rows of modules across
• Gaps within tables for drive
table
systems
• Columns of modules per
• In-table obstructions/
• Tracker with centralized table
protrusions
drive system

13
There are a wide range of cleaning methods and solution
providers – both wet and dry.
Manual cleaning

Semi-automated truck-
mounted cleaning

Semi- and fully-automated


robotic cleaning

Source: Apricum analysis;

14
A large variety of cleaning technology vendors are vying for
business, but only few solutions are effective and bankable.
Cleaning technology vendor summary

Manual devices1 Truck-mounted Semi-automatic Fully automatic


robots robots

Companies

Sub- • Dust broom • Hydraulic arm • Battery powered • Rail mounted


segments • Brush trolley (water, steam or • Cabled • Frame mounted
compressed air)
• Manual water • Free moving
brushes • Pressurize across surface2
water tank and
• Sprinkler2
hose

Source: Apricum analysis based on research and interviews; 1) Not exhaustive, many players; 2) Typically used only for rooftop

15
Certain site or plant design attributes significantly affect
competitiveness for each type of cleaning method.
Factors affecting competitiveness by cleaning method

Manual Truck- Semi-auto Full-auto


mounted robot robotic

+ Low labor cost + Cheap, available + Flexible cleaning + High soiling rates
+ Infrequent water requirements + Water unavailable
cleaning needs + Flexible cleaning + Compatible with + High labor rates
+ Large labor pool requirements tracker systems
+ High predictability
available + Labor can work in + Upfront cost required
cooled vehicles sensitivity
+ Large site
 High inflation rate  Unavailable water  Hot environments  No tracker
 Sensitivity to  Narrow rows (labor exposed) products (yet)
module damage  Difficult terrain  Inaccessible  Sensitivity to
 Inaccessible or water1 upfront cost
 Tracker drive line
expensive water  Obstructions  Obstructions
 Sensitivity to
for wet cleaning within tables within tables
module damage
Source: Apricum analysis; 1) For robots requiring water

16
Manual, truck-based and semi-automatic robotic cleaning
methods require careful tuning of cleaning frequency.
Impact of yield
improvement
Revenues
NPV Ever less
marginal returns
from increasing
cleaning Net project NPV
1x 365x frequency Optimum
Cleaning frequency

Cost Impact of
Very high 26x
NPV O&M cost
cleaning Cleaning frequency
frequency
generates high
OPEX impact
1x 365x
Cleaning frequency

17
Robotic cleaning is much more robust: marginal cost of
cleaning is negligible – high frequency cleaning is optimal.
Impact of yield
improvement
Revenues
NPV Ever less
marginal returns
from increasing
cleaning Net project NPV
1x 365x frequency
Optimum
Cleaning frequency

Cost Impact of 365x


O&M cost Each cleaning
NPV
cycle has Cleaning frequency
negligible
additional cost
1x 365x
Cleaning frequency

18
In regions where labor cost is relatively high, soiling very
high and/or water cost is high, robotic cleaning wins.
Example case: Relative NPV of PV project by cleaning solution [USD cents/Wp]

Dubai
11.6
10.8
8.5
7.4

0.0

Manual – Manual – Water truck Robotic Semi-auto


dust broom brush (dry) robot
(baseline) trolley

Frequency
[days]
25 9 8 1 5

Source: Apricum analysis; assumes fixed tilt system that robotic systems are compatible with

19
Conclusion: cleaning can be a “make or break” factor for
competitive PV projects.
• Cleaning is becoming a crucial topic for
PV developers and IPPs, as projects are
increasingly built in “dirty” regions and
margins for error shrink
• There is no “one size fits all” solution;
methodology and cleaning regime need to
be optimized per site and project layout
• Knowledge of climatic parameters and
careful planning is key!
• Dry robotic technologies promise to be
the most robust and scalable technology
going forward
• Bankability issues still underestimated
• Key challenge for fully-automated robotic solution providers will be to
offer an efficient tracker solution

20
Thank you for your attention

James Kurz
Project Manager
kurz@apricum-group.com
T +49-30-30 87 76 2 29
PV Cleaning: Choosing the Optimal Method and Frequency

MODERATOR
Michael Schmela

Executive Advisor

SPEAKER
James Kurz

Project Manager

Partner
SolarPower Webinar

Thank you for attending

Download this presentation under


PAST WEBINAR on
www.solarpowereurope.org/events/past-events

SolarPower Europe (European Photovoltaic Industry Association)


Rue d’Arlon 69-71, 1040 Brussels, Belgium
Partner T +32 2 709 55 20 / F +32 2 725 32 50
info@solarpowereurope.org /www.solarpowereurope.org

You might also like