Player is loading...
SAMA Renewable Solutions
Wednesday, Oct. 25, 2017 · 11:59 a.m. · 05m 43s
Embed
Transcriptions
Note: this content has been automatically generated.
00:00:00
uh_huh
00:00:05
in summer we have one mission reducing the electric rhythm bonus costs for new ball asset owners
00:00:13
as you probably know with increased proliferation of wind and solar generation assets
00:00:18
is a blowing challenge for these gases owners in managing the risk of
00:00:23
into mittens power production that depends on on certain weather conditions
00:00:28
on one hand the operator requires all market participants to submit
00:00:33
thirty six hours in advance expected power injections the great
00:00:39
on the other hand the wind or solar assets owners
00:00:45
rely on weather forecasts to predict is power injection
00:00:51
however says no for cost is and the percent reliable no
00:00:54
always deviations between the expected and the real output
00:00:59
and these deviations positive or negative are penalise by the
00:01:03
operator i refer to as agreed in bottles costs
00:01:07
which can wipe out a large portion of renewable energy cells right
00:01:13
two addresses challenge would be bought an intelligent algorithm that has three dimensions
00:01:19
the first one it learns from the past deviations in the
00:01:23
four costs in order to correct for the new ones
00:01:28
the second dimension is price
00:01:32
it's models explicitly the market price dynamics which decides whether
00:01:38
is the second largest source of a certainty
00:01:41
and the third dimension uses optimal control to maximise expected
00:01:46
revenues from seven the power on the shorter markets
00:01:50
while minimising the embarrassed cause for new but as it when there's five thirty percent
00:01:57
as you could see in this plot but they profit from
00:02:00
selling the power using our algorithm in the blue line
00:02:04
is always higher than this clients they profits
00:02:09
who uses according to them the best weather forecasts available in the market
00:02:15
and the reason for our performance is that we use the price the mention in the optimisation
00:02:22
currently we have three reading proof of concepts and our business model
00:02:27
is based on the success fees where we charge thirty percent of the cost
00:02:32
saving of the clients can achieve using our software as a service
00:02:39
our team is comprised of four marcus professionals with ninety years combined
00:02:44
experience in mathematics fine as energy trading and i to programming
00:02:49
we worked with each other for many years at r. w. supplied trading the second largest initiatives in europe
00:02:54
where i had to the quantitative analysis departments and patterson young when i was an exacting director
00:03:02
we are raising it haven't fifty thousand swiss francs to expand our markets prizes and to re
00:03:09
architect our software the service to become production really be part of our success thank you
00:03:15
hi grandma she and and actually ah
00:03:25
i'm i'm i'm i i had this
00:03:31
yes uh this is the uh results of five years of research and development
00:03:37
uh it is a notoriously very complex problem we surveyed twelve whether for
00:03:42
cost providers to compare the wind into power and twenty and utilities
00:03:48
none of them has both have both the price and the weather combined in the optimisation
00:03:54
it is a complex problem because it calls for a combined skill in i. t. mouse or purses research and so
00:04:04
no it is a trade secrets
00:04:13
exact no daily on the basis we push sorry or hourly bases we push
00:04:18
our forecasts and compare it hand in hand with their own for costs
00:04:23
i mean we measure as you could see in this graph is is for a pretty particular
00:04:27
period how much p. n. l. or profit to make compared to their own internal mops
00:04:34
it's a very long city life cycle is our initiatives is like
00:04:38
x. p. one on peak where have lot of uh priorities
00:04:43
this is very important for them but you know to assess the quality we need to have
00:04:47
a t. v. in place to check the numbers and that's talks takes a long time
00:04:54
we try not to do to my daughter's defiant have a tan um companies in our pipeline to
00:04:59
pursue the efforts of saying one pile so the first what comes first and and so on
00:05:06
upon one another question ah yes
00:05:13
we tried three that us forces uh the best when it could
00:05:16
find is in switzerland uh in is a company here uh
00:05:21
the idea is to the weather for cause is only one input
00:05:25
it is important but this is not the main a driver for the higher piano but we
00:05:31
are we tried to read data sources and all of that is was and best
00:05:37
okay thank you bye and that ah
Conference Program
SAMA Renewable Solutions
Oct. 25, 2017 · 11:59 a.m.
IntelliCred
Oct. 25, 2017 · 12:35 p.m.
115 views
Recommended talks
How to align the digital with the ecological transition
Lorenz M. Hilty, Professor of Informatics and Sustainability at the University of Zurich
Feb. 12, 2016 · 11:41 a.m.
162 views