satellite- and (camera-derived) irradiance data for ... · satellite- and (camera-derived)...

Deutsches Fernerkundungsdatenzentrum

Satellite- and (camera-derived) irradiance data for applications in low voltage grids with large PV shares

Marion Schroedter-Homscheidt, Gerhard Gesell, Lars Klüser, Miriam Kosmale, Sandra Jung, Niels Killius

Deutsches Zentrum für Luft- und Raumfahrt (DLR)

Earth Observation Center (EOC)

Holger Ruf, Gerd Heilscher Hochschule Ulm Contributions from V.H. Peuch and A. Benedetti (ECMWF), L. Wald and M. Lefèvre (Armines), E. Wey (Transvalor)


Introduction

• Satellite-based solar resource information – New Copernicus/MACC service

• What‘s that? What is new?

• What else do we with the intermediate Copernicus service results?

• Cloud/snow statistics

• Dust aerosol statistics

• Nowcasting clouds for large solar power plants -> low voltage grids?

• Feed-in power and transformer load flow in a low voltage grid

• Cloud cameras – can we use NWP instead of a ceilometer for cloud height assessment?


Introduction

• Satellite-based solar resource information – New Copernicus/MACC service

• What‘s that? What is new?

• What else do we with the intermediate Copernicus service results?

• Cloud/snow statistics

• Dust aerosol statistics

• Nowcasting clouds for large solar power plants

• Feed-in power and transformer load flow in a low voltage grid

• Cloud cameras – can we use NWP instead of a ceilometer for cloud height assessment?


MACC-III - The Copernicus Atmosphere Service‘s precursor

Air quality

Global Pollution

Aerosol UV

http://atmosphere.copernicus.eu


Physical retrieval method Heliosat-4

Scattering through

clouds and

aerosols

diffuse

irradiance

direct and diffuse

irradiance

Global irradiance

Absorption from

clouds,

aerosols, and

water vapour

diffuse

irradiance

Direct

normal

irradiance

Fast radiative transfer model = Heliosat-4


McClear clear sky irradiance (GHI, DIF, DIR, DNI) time series in 1 min steps and in global coverage

• 2004-2015 1-2 days delay online

• 7 years operations secured,

> 20 years planned

• Just register and

download from

http://www.soda-pro.com/ web-

services/radiation/

mcclear

• Interactive and OGC script

access possible


MACC-RAD: Heliosat-4 irradiance (GHI, DIF, DIR, DNI) time series in 15 min steps in Europe/Africa/Middle East

• 2004-2015 1-2 days delay online

• 7 years operations secured,

> 20 years planned

• Just register and

download from

http://www.soda-pro.com /web-

services/radiation/

macc-rad

• Interactive and OGC script

access possible


Registration needed to justify existence of service

Free for any use – commercial as well as R&D

Only restriction: Do not sell the data itself without modifications

Data policy

Here you can find Data Policy and

User‘s Guide


Olympic goals in solar resources: Aim to be more accurate, better resolved, more parameters…

But is that all?

Today’s situation: ‘I’m asking 5 data providers and will get 6 different answers … and now there is also MACC/Copernicus’

Sure, better, more accurate, more parameters is needed…

But also:

transparency – which input data is used? Publish instead of company confidential status

Detailed validation – publish instead of treated as company confidential

Continuous validation and quality control with experienced staff – no project like approach with new PostDoc students every few years

Faster, higher, wider,…. Is that all? Or: Why a new Copernicus service?


What can we do with Copernicus/MACC intermediate data?


Cloud and snow statistics

Carpentras, France

green = clear

blue = overcast/broken clouds

yellow = cirrus, thin ice

red = scattered clouds


Cloud and snow statistics

green = clear

blue = overcast/broken clouds

yellow = cirrus, thin ice

red = scattered clouds

La Réunion

Carpentras, France


Dust aerosol statistics – extinction, dust events from MATCH

GHI, IEED_5 GHI, IEED_10

number of exceedance days per year

IEED = Irradiance Extinction Exceedance Day Thresholds 5, 10, 30% for GHI 10,30,50,70% for direct irradiances Based on a 30 years climatology of daily dust aerosol optical depth


Duration of dust events – as seen in MATCH

maximum number of consecutive IEED days

GHI, IEED for 10% extinction

mean number of consecutive IEED days

GHI, IEED for 10% extinction


Nowcasting for solar power plants

©

EUMETSAT/DLR

sequence of cloud masks

thin ice clouds

water/mixed phase clouds

cloud mask

forecast

irradiance forecast (Heliosat-4)

©

EUMETSAT/DLR

©

EUMETSAT/DLR

production forecast

market operations

© Eumetsat/DLR


Feed-in power and transformer load – Ulm example

Grid simulation

Solar irradiance

at the surface

monitoring

forecast

Optimized solar energy integration

into the existing grid – grid status (historical data)

and nowcasting (expected to be used)

NWP

-17

-6

-45

-3/1

-11

-8

-5 -5

-69

-25

-26 -18 -24

-11

-2 -4

-58

-3

-67


Pyranometers

more accurate at the point

few locations only

maintenance needed

Satellite

spatially averaged

15 or 5 minutes

Long history available

Satellite or ground observations ?


Test site Ulm Einsingen

• 12 PV-Systems with smart meters (red marked)

• 15 minute resolution

• 6.7.2013 – 13.12.2014

• Comparison of individual systems and sum


Feed-in power – Comparison vs smart meters

Pyranometer (HSU) MACC-RAD


From feed-in power to load flow at the transformer


Load flow at transformer

Only values with sun elevation angles >0° are considered

Annual results for 2012


Sky cameras and cloud height – can we save money for a ground based ceilometer network and use the numerical weather prediction instead ?

Besides the question if sky cameras with their 10 minute forecast horizon and restricted field of view are of interest at all for low voltage systems… , but PV and CSP power plants might be interested…

A bit on cameras…


Cameras and cloud height

Cloud detection

Cloud position

Cloud motion

Nowcast

y

x

z


Remember – ceilometers are not cheap and they are point obs


A real case in a ceilometer at PSA, Almeria, Spain

within 5

minutes grey =

P25 and P75

within 5 min

MODIS image

at 13:10 UTC


Can I use numerical weather prediction‘s cloud condensation levels as from ECMWF?

multilayer

situation

All cases scattered

Rmsd 5.5 km 5.5 km

Bias 1.09 km 1.17 km

Pearson

corr.

0.50 0.43

only scattered

all


The new MACC/Copernicus irradiance service has been presented

Usage of cloud properties from space for extended site characteristics

Usage of aerosol properties from long-term modeling for site characteristics

Cloud properties used in nowcasting for large scale solar power plants

MACC irradiances for feed-in power calculation and transformer loads in low voltage grids

Quantification of ECMWF cloud base height vs ceilometer in Spain

Wrap-Up


This work has recently received funding from the European Union’s Seventh Programme for research, technological development and demonstration under grant agreement No 608623 and 608930.

Thanks to … … Stadtwerke Ulm for providing smart meter and transformer data … the MACC team for providing data … ECMWF for providing forecast data … EUMETSAT for providing MSG satellite observations

satellite- and (camera-derived) irradiance data for ... · satellite- and (camera-derived)...

Documents