Take Control Of Your Electrical Systems With The Open Source FlexMeasures Energy Management System

WEBVTT
NOTE
Transcription provided by Podhome.fm
Created: 7/6/2024 4:55:13 PM
Duration: 2956.844
Channels: 1
1
00:00:13.344 --> 00:00:24.700
Hello, and welcome to podcast dot in it, the podcast about Python and the people who make it great. When you're ready to launch your next app or want to try a project you hear about on the show, you'll need somewhere to deploy it. So check out our friends over at Linode.
2
00:00:25.295 --> 00:00:32.595
With their managed Kubernetes platform, it's easy to get started with the next generation of deployment and scaling powered by the battle tested Linode platform,
3
00:00:32.960 --> 00:00:38.980
including simple pricing, node balancers, 40 gigabit networking, and dedicated CPU and GPU instances.
4
00:00:39.815 --> 00:00:47.355
And now you can launch a managed MySQL, Postgres, or Mongo database cluster in minutes to keep your critical data safe with automated backups and failover.
5
00:00:47.660 --> 00:00:48.800
Go to python podcast.com/linode
6
00:00:50.220 --> 00:00:56.640
today to get a $100 credit to try out their new database service, and don't forget to thank them for their continued support of this show.
7
00:00:57.085 --> 00:01:04.545
The biggest challenge with modern data systems is understanding what data you have, where it is located, and who is using it. SelectStar's
8
00:01:04.845 --> 00:01:15.229
data discovery platform solves that out of the box with a fully automated catalog that platform solves that out of the box with a fully automated catalog that includes lineage from where the data originated all the way to which dashboards rely on it and who is viewing them every day.
9
00:01:15.585 --> 00:01:17.125
Just connect it to your DBT,
10
00:01:17.425 --> 00:01:23.205
Snowflake, Tableau, Looker, or whatever you're using, and select star will set everything up in just a few hours.
11
00:01:23.619 --> 00:01:24.920
Go to python podcast.com/selectstar
12
00:01:26.579 --> 00:01:31.240
today to double the length of your free trial and get a swag package when you convert to a paid plan.
13
00:01:31.585 --> 00:01:40.200
Your host as usual is Tobias Macy. And today, I'm interviewing Nicholas Hoehning about Flex Measures, an open source project designed to manage energy resources dynamically
14
00:01:41.880 --> 00:01:43.820
Nicholas, can you start by introducing yourself?
15
00:01:44.280 --> 00:01:47.740
Yes. Hi, Tobias. I'm Nicholas. I'm German
16
00:01:48.905 --> 00:01:50.125
and living in the Netherlands
17
00:01:50.664 --> 00:01:55.485
since 15 years. I've been a programmer for a bit more than 20 years
18
00:01:56.020 --> 00:02:00.680
using Python for a long time. I've been working on energy since 12 years, I think,
19
00:02:01.060 --> 00:02:02.680
from that vintage point.
20
00:02:03.355 --> 00:02:05.615
And do you remember how you first got introduced to Python?
21
00:02:06.155 --> 00:02:08.015
1st, I was I had been using
22
00:02:09.355 --> 00:02:13.030
very little c and Visual Basic. I've been working on
23
00:02:13.489 --> 00:02:13.989
access
24
00:02:14.370 --> 00:02:14.870
databases
25
00:02:15.170 --> 00:02:16.310
and and word
26
00:02:17.010 --> 00:02:17.510
macros.
27
00:02:18.185 --> 00:02:20.845
At at some point, I did then, go to university.
28
00:02:21.705 --> 00:02:26.519
During my bachelor's, I got to work on a start up. It was a neuroscience
29
00:02:26.980 --> 00:02:27.480
startup
30
00:02:28.500 --> 00:02:30.680
that is still around. It's called Iquant.
31
00:02:30.980 --> 00:02:32.360
The idea was to
32
00:02:32.875 --> 00:02:33.534
transfer neuroscience
33
00:02:34.635 --> 00:02:36.174
models of visual attention
34
00:02:36.715 --> 00:02:40.735
into a web application so that you can, you know, basically upload your
35
00:02:41.194 --> 00:02:42.415
ad and find out
36
00:02:42.780 --> 00:02:44.159
where people might look first.
37
00:02:44.459 --> 00:02:54.594
And we do we use a lot of Python there, and I had to you know, I was sitting at lunch, reading the standard Python tutorial so I could get to work on that startup, still remember.
38
00:02:55.055 --> 00:03:06.160
In terms of the Flex Measures project that you are working with, can you describe a bit about what it is and some of the story behind how it came to be and why you decided that this dynamic
39
00:03:06.575 --> 00:03:10.755
energy demand response system is something that you wanted to spend your time and energy on?
40
00:03:11.135 --> 00:03:14.355
Yeah. Sure. And I'd love to yeah. It has become my life topic.
41
00:03:14.770 --> 00:03:16.070
I started a PhD
42
00:03:16.450 --> 00:03:17.910
around this topic in 2009
43
00:03:19.090 --> 00:03:19.590
when
44
00:03:20.450 --> 00:03:27.445
it wasn't absolutely clear that we will get the kind of energy transition that we are basically in right now,
45
00:03:28.145 --> 00:03:30.325
driven by, you know, lots of renewables
46
00:03:31.450 --> 00:03:33.710
being the major source at some point.
47
00:03:34.250 --> 00:03:36.910
Flex measures, basically, an energy management system.
48
00:03:37.610 --> 00:03:42.385
It helps you manage your energy. In the last couple of years, energy management has meant that you manage
49
00:03:42.924 --> 00:03:45.344
the data around your energy, metering data.
50
00:03:46.045 --> 00:03:54.750
You put it on NICE dashboards and you do billing around it. You know, that's also significant work to be done. Flex Measures focuses more on
51
00:03:55.130 --> 00:03:56.430
the energy flexibility
52
00:03:57.915 --> 00:03:58.655
and optimizing
53
00:03:59.355 --> 00:04:02.015
when energy is consumed during the day
54
00:04:02.475 --> 00:04:05.535
because that's an a very important piece in the energy transition.
55
00:04:06.690 --> 00:04:11.270
As any any expert will tell you, we'll need to get to this kind of optimization,
56
00:04:12.130 --> 00:04:13.030
at some point.
57
00:04:13.865 --> 00:04:16.685
Otherwise, we will not get the fit between
58
00:04:17.465 --> 00:04:25.159
available renewable energy and the actual consumption to have a stable energy system. So flex measures helps you find the best times
59
00:04:25.860 --> 00:04:31.960
for your flexible energy assets. Let's say, a battery. That's the best example of a flexible energy asset.
60
00:04:32.365 --> 00:04:34.945
When should that battery be charging or discharging?
61
00:04:35.405 --> 00:04:40.544
The other examples for flexible assets are heating. You can heat before
62
00:04:41.229 --> 00:04:43.650
somewhat sometime before you're really home, for instance.
63
00:04:44.110 --> 00:04:45.569
Heat pumps are a big topic.
64
00:04:46.110 --> 00:04:48.050
The industry, you have a lot of potential
65
00:04:48.815 --> 00:04:54.035
to run energy intensive processes at a slightly different time. And Flex Measures really helps you
66
00:04:54.495 --> 00:04:57.875
model all your energy data in real time and make that decision.
67
00:04:58.380 --> 00:05:26.794
You mentioned that a number of the other energy management systems are focused more on a kind of 1 way path of information where you're just retrieving data from the system and then using that to build some analysis, but that doesn't get acted upon in any sort of automated fashion. I'm wondering if you can just give a bit more of an overview of what the space looks like for these energy management systems and if there are any kind of alternatives to flex measures or if it's something that is relatively unique in the space.
68
00:05:27.255 --> 00:05:28.315
There are some projects
69
00:05:28.680 --> 00:05:32.220
that have on the road map to also do this kind of real time optimization,
70
00:05:33.160 --> 00:05:38.060
but you got a lot of inertia in these projects as well in companies. So I see a lot of companies who
71
00:05:39.045 --> 00:05:43.305
had to start basically going to customers who have not even any metering
72
00:05:43.925 --> 00:05:44.985
installed whatsoever
73
00:05:45.605 --> 00:05:49.330
other than having a main meter. And the main meter's data
74
00:05:50.190 --> 00:05:51.810
is not real time as well.
75
00:05:52.350 --> 00:05:56.370
So these companies started with bringing in their Raspberry Pis or whatever it is,
76
00:05:56.885 --> 00:06:04.905
becoming good at that, starting to measure, becoming good at visualizing information, becoming good at doing them some some aspects of billing maybe.
77
00:06:05.720 --> 00:06:09.340
And to do this step into data science optimization,
78
00:06:09.720 --> 00:06:10.380
I see
79
00:06:10.920 --> 00:06:14.620
I see the difficulties in some organizations to to that extra step. So
80
00:06:14.965 --> 00:06:16.025
these people are happy
81
00:06:16.965 --> 00:06:27.520
to get some other project and get some other help with that. I don't know all the projects out there, of course. So Flex Measures is open source, and another open source project that I know about is just called OpenEMS.
82
00:06:28.540 --> 00:06:29.740
That's a Java project in,
83
00:06:30.905 --> 00:06:32.125
originated in Germany
84
00:06:32.425 --> 00:06:39.645
that's, built a quite interesting foundation around it. But also there, there's a big focus in inertia on getting the metering
85
00:06:40.449 --> 00:06:41.110
and visualization
86
00:06:42.289 --> 00:06:43.190
correctly done.
87
00:06:43.650 --> 00:06:46.710
And there's not so much support yet to
88
00:06:47.625 --> 00:06:48.745
optimize the actual
89
00:06:49.785 --> 00:06:54.045
what actually happens and and influence that, and we have decided to only focus on that.
90
00:06:54.639 --> 00:06:55.780
As far as the
91
00:06:56.560 --> 00:07:01.965
primary focus and goals of Flex Measures, I'm wondering if you can just talk to some of
92
00:07:02.445 --> 00:07:04.945
the kind of outcomes that you're aiming for and
93
00:07:05.565 --> 00:07:06.465
some of the,
94
00:07:06.925 --> 00:07:09.185
I guess, target audience for the project.
95
00:07:09.780 --> 00:07:14.680
As a big goal, the big hairy goal is that Plaxminder should really help to
96
00:07:15.060 --> 00:07:20.815
speed up the energy transition and, you know, become 1 of the standard pieces of climate tech,
97
00:07:21.755 --> 00:07:26.095
infrastructure tech that you would also always consider when you start a new
98
00:07:26.840 --> 00:07:28.700
modern energy project where
99
00:07:29.240 --> 00:07:30.540
your real time optimization
100
00:07:31.320 --> 00:07:32.780
is a part of the solution.
101
00:07:33.535 --> 00:07:37.635
The main way I think Flex Measures can help is really being a developer friendly tool,
102
00:07:38.255 --> 00:07:39.395
solve a lot of these
103
00:07:39.775 --> 00:07:44.830
things that take you weeks of work, get your data model right, get your data integrations right,
104
00:07:45.210 --> 00:07:46.350
get some monitoring
105
00:07:47.130 --> 00:07:47.630
right,
106
00:07:48.090 --> 00:07:51.550
decide how to implement where your algorithms should be run,
107
00:07:51.965 --> 00:07:54.305
have a queue for that, you know, processing queues.
108
00:07:55.085 --> 00:07:57.585
There's all these things that developers know
109
00:07:58.125 --> 00:07:59.105
will keep you
110
00:07:59.500 --> 00:08:01.200
busy before you get to the real
111
00:08:01.580 --> 00:08:06.000
use case, before you really help your customer, or you are implementing something very
112
00:08:06.540 --> 00:08:08.560
shortly, very quick hack.
113
00:08:08.875 --> 00:08:16.415
See that happening a lot. Other people asking their cousin's neighbor to have a simple algorithm put somewhere, but this is about, you know,
114
00:08:16.795 --> 00:08:26.960
real time data pipelines who have to be kept alive and be checked on health. Yeah. I think we succeed if developers around the world who I expect in the next 10 years to
115
00:08:27.345 --> 00:08:33.445
be looking around how they should approach the similar similar challenges to always consider flex measures.
116
00:08:33.825 --> 00:08:35.765
And that can be in a start up somewhere.
117
00:08:36.430 --> 00:08:37.570
That can be in
118
00:08:38.029 --> 00:08:39.490
a larger energy company
119
00:08:39.950 --> 00:08:45.810
where they have already their ICT team, but, haven't solved or started to do this yet.
120
00:08:47.105 --> 00:08:48.085
And and I think
121
00:08:48.785 --> 00:08:53.445
this is a problem that happens in all the the the parts of the energy system. For me,
122
00:08:54.000 --> 00:08:56.820
there's 2 big parts. That's industry and,
123
00:08:57.200 --> 00:08:58.019
what's usually
124
00:08:58.480 --> 00:09:07.405
called smart city at the moment. So anything in the the built environment, those are 2 huge sectors of the energy sector, which in itself is 1 of the biggest
125
00:09:08.240 --> 00:09:16.100
sectors in the world that needs to undergo the most rapid transition you could imagine. The other plans are 15 year plans basically to basically change
126
00:09:16.800 --> 00:09:17.940
how everything works.
127
00:09:18.515 --> 00:09:23.495
So we really hope flex measures can help get that super big vision done in time.
128
00:09:24.035 --> 00:09:35.970
Do you also see flex measures as something that an individual homeowner might use as part of managing their energy systems, particularly as more people invest in solar and small scale wind?
129
00:09:36.415 --> 00:09:40.595
Yes. I want to talk about a small project we've been doing. So
130
00:09:41.135 --> 00:09:46.320
besides working with customers or with companies that can scale up innovation.
131
00:09:47.020 --> 00:09:59.214
We've seen that this is still slow. Right? So these companies all have to think differently before they haven't been involved with energy so far. It's a long process. So we have, on the sides, created a living lab with just 1 household
132
00:09:59.675 --> 00:10:01.615
with someone who's also very enthusiastic
133
00:10:02.730 --> 00:10:03.550
about, specifically,
134
00:10:04.410 --> 00:10:05.709
vehicle to grid charging.
135
00:10:06.170 --> 00:10:12.029
They bought the exact car and the exact charger that can do that already, the 1 that is allowed to do that in the Netherlands.
136
00:10:12.394 --> 00:10:13.774
And we've really implemented
137
00:10:14.235 --> 00:10:15.615
the automated optimization
138
00:10:16.235 --> 00:10:22.095
of the charging and discharging in in his house. And for that, we needed to actually do something in the house.
139
00:10:22.459 --> 00:10:25.279
Flex measures is basically supposed to run-in the cloud.
140
00:10:25.819 --> 00:10:27.839
It's dockerized. It can also run locally.
141
00:10:28.379 --> 00:10:31.920
But it basically makes decisions. It doesn't really show you something.
142
00:10:32.325 --> 00:10:36.265
So I want to pitch a very cool project that's called Home Assistant.
143
00:10:36.645 --> 00:10:39.785
It's also open source. And it's also originated in the Netherlands,
144
00:10:40.245 --> 00:10:41.385
but they have a
145
00:10:41.899 --> 00:10:44.480
company in the US also now. Very interesting
146
00:10:44.940 --> 00:10:47.279
group of people. Basically, in the home automation
147
00:10:47.820 --> 00:10:49.925
sector with their software.
148
00:10:50.465 --> 00:11:02.940
We use Home Assistant together with Flex Measures to optimize the charging, discharging, and and taking into account the the solar panel production to really create a cool project. We also open sourced our Home Assistant plug in.
149
00:11:03.560 --> 00:11:14.825
So that's out there. We were on Dutch TV and a couple of people wanted to use it as well. And now there's this week, the first 2 others who've implemented it in their house. So now we have 3.
150
00:11:15.320 --> 00:11:22.300
But for us, as a company, that's not the focus, but it's where you can really do something right now and really show something. And there's always these, enthusiasts
151
00:11:22.925 --> 00:11:25.504
that just want to build something in their house like that.
152
00:11:25.805 --> 00:11:28.144
As far as the overall question
153
00:11:29.004 --> 00:11:31.165
of energy systems, I mean, that can be
154
00:11:31.920 --> 00:11:34.660
it's a very broad term with a lot of different,
155
00:11:35.360 --> 00:11:38.500
I guess, energy sources and use cases
156
00:11:39.105 --> 00:11:48.325
kind of embedded in it, where 1 element of it is the electrical grid. Another is the usage of fossil fuels for transportation and shipping.
157
00:11:48.650 --> 00:11:53.630
There's home heating. There's heating at, you know, industrial scale for large factories.
158
00:11:54.410 --> 00:11:56.510
And the majority of those,
159
00:11:56.925 --> 00:12:17.505
I guess, ecosystems of energy usage are governed largely by different market forces where there is the, you know, energy producer, which is becoming a bit of a fuzzy line now with more homes and residential and industrial installations having their own energy generation through things like solar and wind. And then there's the energy consumer and the distribution system.
160
00:12:18.045 --> 00:12:28.800
And the pricing of those different resources are governed by, you know, some level of dynamic usage where, you know, there are peak loads, and so that can increase energy prices for certain periods.
161
00:12:29.339 --> 00:12:30.160
And I'm wondering
162
00:12:30.644 --> 00:12:34.985
what you see as some of the benefits that Flex Measures can provide to
163
00:12:35.365 --> 00:12:41.339
this marketplace ecosystem, particularly given the fact that it's not as clear cut now as it once was?
164
00:12:41.720 --> 00:12:43.740
No. You're absolutely right. It's the vast,
165
00:12:44.120 --> 00:12:45.100
sometimes confusing
166
00:12:45.880 --> 00:12:54.115
system. I like the analogy that because somehow it's all connected and it has to happen, in real time and the frequency of the grid,
167
00:12:54.580 --> 00:12:58.040
as far as electricity is concerned is held in some
168
00:12:58.660 --> 00:12:59.160
synchronization.
169
00:12:59.940 --> 00:13:13.200
We've created the largest man made machines with the grids. So you got the US grid, the European grid, and the Asian grid, basically, as the largest 3 machines you could even imagine where your house is a part of it, or the electric assets,
170
00:13:13.820 --> 00:13:16.399
and it happens in real time. Different markets
171
00:13:17.180 --> 00:13:21.040
buy and sell and allocate energy for different horizons.
172
00:13:21.915 --> 00:13:24.495
And FlexMaster is certainly
173
00:13:25.195 --> 00:13:31.375
not very useful in the range of, let's say, you're buying your energy a couple months ahead and buying these big budgets.
174
00:13:32.140 --> 00:13:41.415
Also, I wouldn't say it's usable in the second range. There's some balancing markets to get this frequency challenge on the grid, right, where you have to react within a second.
175
00:13:41.954 --> 00:13:43.175
But everything in between,
176
00:13:43.714 --> 00:13:46.855
where you have to react basically during the day
177
00:13:47.240 --> 00:13:50.940
or you think in the evening about your plan the next day,
178
00:13:51.400 --> 00:13:52.860
that's where Flex Measures,
179
00:13:53.320 --> 00:13:54.220
is really useful.
180
00:13:54.600 --> 00:13:57.324
And as I said, it's really on the electricity
181
00:13:58.105 --> 00:13:58.605
range
182
00:13:59.065 --> 00:14:00.605
where we want to use it.
183
00:14:01.065 --> 00:14:02.444
So if you're a consumer,
184
00:14:02.959 --> 00:14:14.785
you can use flex measures to use greener electricity, but it will help the grid stability. But I think mostly what people care about, of course, is to to have their energy bill in check. The spread on market prices has gone up.
185
00:14:15.245 --> 00:14:16.145
Energy bills
186
00:14:16.685 --> 00:14:20.305
are rising a lot. So there's a lot of urgency, I think. Now there's
187
00:14:20.720 --> 00:14:23.860
clarity that this is very important for them to achieve.
188
00:14:24.240 --> 00:14:24.740
And
189
00:14:25.120 --> 00:14:26.260
from a grid perspective,
190
00:14:26.880 --> 00:14:28.180
if you're a network operator,
191
00:14:28.875 --> 00:14:31.295
it also is great if a lot of
192
00:14:31.675 --> 00:14:33.695
decentralized decisions are happening
193
00:14:34.154 --> 00:14:35.695
that help the grid stability.
194
00:14:36.240 --> 00:14:40.820
We we have to follow what renewables are doing to a big extent more and more.
195
00:14:41.200 --> 00:14:45.845
We cannot influence when we get that solar power or wind power, but we can influence
196
00:14:47.024 --> 00:14:49.524
when we use power. And if a lot of decentralized
197
00:14:49.985 --> 00:14:52.404
decisions are being made to shift their consumptions
198
00:14:52.879 --> 00:14:56.819
towards when renewable energy is really there. We get a more stable grid.
199
00:14:57.120 --> 00:14:57.620
And
200
00:14:58.000 --> 00:15:03.415
the only other solution for network operators would be to, you know, double or triple every cable in size.
201
00:15:03.955 --> 00:15:04.455
So
202
00:15:04.835 --> 00:15:06.055
this is a big societal
203
00:15:06.435 --> 00:15:08.695
cost saving we can achieve with installing
204
00:15:09.420 --> 00:15:19.985
better decision making in in the nodes, which turns how the energy system works on its head. Right? So still now in the last decades, we have optimized a few power plants per country
205
00:15:20.365 --> 00:15:23.185
and just forecasted what consumers will be doing.
206
00:15:23.565 --> 00:15:26.704
And now that partly turns on its head. It's very interesting.
207
00:15:27.110 --> 00:15:30.410
Digging a bit more into that question of renewables
208
00:15:30.870 --> 00:15:37.155
and their broader distribution, particularly with residential installations as part of that mix, I'm curious
209
00:15:37.695 --> 00:15:39.315
what you see as some of the
210
00:15:39.695 --> 00:15:45.795
broader impact on the way that the kind of marketplace economy around electrical grids in particular
211
00:15:46.480 --> 00:15:50.980
is likely to shift as the power generation and power distribution
212
00:15:51.440 --> 00:15:51.940
becomes
213
00:15:52.320 --> 00:15:52.820
bidirectional
214
00:15:53.360 --> 00:15:54.980
on a more routine basis.
215
00:15:55.535 --> 00:16:01.315
1 thing we see in the Netherlands, and I think that the Netherlands are a bit ahead of the curve there, is
216
00:16:01.774 --> 00:16:02.995
more dynamic prices.
217
00:16:03.500 --> 00:16:04.959
I think you can choose between
218
00:16:05.660 --> 00:16:08.560
5 suppliers now that will buy
219
00:16:09.019 --> 00:16:11.839
power on the day ahead market, which has 24
220
00:16:12.754 --> 00:16:13.254
different
221
00:16:13.795 --> 00:16:15.254
slots. But then they will just
222
00:16:15.634 --> 00:16:16.615
give you those
223
00:16:16.995 --> 00:16:18.055
24 prices.
224
00:16:18.595 --> 00:16:20.055
So the the the afternoon
225
00:16:20.920 --> 00:16:25.899
before, you will know what your 24 prices will be the next day. And if you have,
226
00:16:26.279 --> 00:16:27.259
you know, some optimized
227
00:16:28.145 --> 00:16:32.325
optimized system like what we did, for instance, in the vehicle to grid project,
228
00:16:32.705 --> 00:16:39.320
that's going to turn out quite well for you then. You will not consume energy when those prices are high. Of course,
229
00:16:39.940 --> 00:16:45.000
what's interesting is that this might create more of a divide between those people that can afford that
230
00:16:45.505 --> 00:16:46.805
that has this cool
231
00:16:47.265 --> 00:16:51.605
infrastructure plus software, and they get their average energy bill.
232
00:16:51.985 --> 00:16:53.700
Well, they keep it low, let's say.
233
00:16:54.180 --> 00:16:59.800
And who's going to pay a lot? That's probably going to if everything goes like this, it's going to be the poor people.
234
00:17:00.420 --> 00:17:02.360
So this is outside of the technical
235
00:17:03.145 --> 00:17:04.605
realm, also very interesting
236
00:17:05.385 --> 00:17:12.045
political discussion where we still have some time. And I'm also involved in some of these discussions to discuss how we can
237
00:17:12.540 --> 00:17:13.040
achieve
238
00:17:13.580 --> 00:17:15.360
a more just way to do that.
239
00:17:15.740 --> 00:17:18.880
What we really need is not for individual homeowners to
240
00:17:19.635 --> 00:17:21.895
invest in their own house and make it smarter.
241
00:17:22.435 --> 00:17:27.895
I like it. I'm enthusiastic about it when it happens, but what we really need is this scale.
242
00:17:28.250 --> 00:17:31.309
We really need to do these things for everybody on a big scale.
243
00:17:31.850 --> 00:17:33.470
So for instance, big housing
244
00:17:33.850 --> 00:17:34.350
corporations
245
00:17:35.455 --> 00:17:37.475
already know, given their ESG
246
00:17:38.255 --> 00:17:38.755
strategy,
247
00:17:39.535 --> 00:17:40.995
they have to renovate
248
00:17:41.375 --> 00:17:50.350
at least 4% of the whole portfolio every year. And every year that they don't do that, it's just getting to it, getting to be more pressing, like,
249
00:17:50.650 --> 00:17:51.790
5%, 6%.
250
00:17:52.490 --> 00:17:54.190
And then, yeah, you need to
251
00:17:54.684 --> 00:17:57.585
make a whole neighborhood or a whole big flat
252
00:17:58.445 --> 00:18:00.144
react to flexibility
253
00:18:00.684 --> 00:18:01.904
demands from the grid.
254
00:18:02.260 --> 00:18:09.560
And that can be very exciting that you then, for instance, have 1 giant heat pump for a whole flat. And especially if you get to bigger numbers,
255
00:18:10.180 --> 00:18:11.080
there will be
256
00:18:11.524 --> 00:18:12.904
a better way to react.
257
00:18:13.524 --> 00:18:15.304
For instance, it's almost impossible
258
00:18:15.605 --> 00:18:19.705
to forecast for a single house when you will be home, when will you cook.
259
00:18:20.120 --> 00:18:25.820
You can have a little bit of a idea. But if you forecast a 100 consumers,
260
00:18:26.760 --> 00:18:28.620
you get a much better data
261
00:18:29.184 --> 00:18:32.005
forecast, which you can really work with. So I think
262
00:18:32.465 --> 00:18:36.005
that's really what I need to see, these these bigger scaled approaches.
263
00:18:36.590 --> 00:18:39.890
And, yeah, I hope Flex Measures can be can be helpful with that.
264
00:18:40.750 --> 00:18:43.570
Digging into some of the way that Flex Measures
265
00:18:43.975 --> 00:18:51.434
operates, I'm wondering if you can talk to the sources of data that you're looking at, some of the ways that you manage that integration
266
00:18:51.815 --> 00:18:52.875
and processing,
267
00:18:53.690 --> 00:19:03.915
and the types of interaction that different kind of users or operators might have with Flex Measures and how much of it is just you kind of set it and forget it? Yes. Sure.
268
00:19:04.615 --> 00:19:08.795
So my simple picture of flex measures is data in, scheduling out.
269
00:19:09.415 --> 00:19:11.915
Data in means certainly metering data.
270
00:19:12.269 --> 00:19:14.129
You want to know what all your assets
271
00:19:14.750 --> 00:19:22.215
did in the past. It can be very useful to, for instance, for a battery, know the possible state of charge it has now.
272
00:19:22.755 --> 00:19:26.135
It can help to forecast what will happen in the near future.
273
00:19:26.674 --> 00:19:28.294
We also read in prices,
274
00:19:29.170 --> 00:19:31.190
market prices, for instance, the next day,
275
00:19:31.890 --> 00:19:32.790
weather forecasts.
276
00:19:33.970 --> 00:19:38.230
So you can then forecast your solar generation if that's on your roof.
277
00:19:38.835 --> 00:19:40.055
Sometimes the occupants'
278
00:19:40.595 --> 00:19:46.775
plans are also data that you need to take in. So in the vehicle to grid, living lab in this application,
279
00:19:47.919 --> 00:19:53.460
The occupants' data might be that they say they will need the car for a longer trip
280
00:19:54.000 --> 00:19:57.455
tonight at 6 PM. That's important data that influences
281
00:19:57.995 --> 00:20:01.135
the scheduling because the car should be a 100% full by then.
282
00:20:01.995 --> 00:20:03.215
Or in a factory,
283
00:20:03.870 --> 00:20:07.330
they also have new constraints every day, how much
284
00:20:07.790 --> 00:20:09.730
processing work they need to be doing.
285
00:20:10.110 --> 00:20:10.590
That's,
286
00:20:10.990 --> 00:20:15.645
about the data. And you said earlier that set and forget might be
287
00:20:16.025 --> 00:20:17.965
an interesting concept. I really agree.
288
00:20:18.585 --> 00:20:18.905
For,
289
00:20:19.465 --> 00:20:20.365
most people
290
00:20:21.250 --> 00:20:23.830
who will live in modern energy environments
291
00:20:24.289 --> 00:20:25.429
with with some optimization
292
00:20:25.809 --> 00:20:26.309
happening,
293
00:20:26.929 --> 00:20:28.150
they should not even
294
00:20:28.925 --> 00:20:31.105
care about that during a normal day.
295
00:20:31.405 --> 00:20:37.105
So in the vehicle to grid project that we've been doing, that really was also design goal.
296
00:20:37.805 --> 00:20:38.545
It worked.
297
00:20:38.910 --> 00:20:40.930
And from then on, it just quietly
298
00:20:41.390 --> 00:20:43.730
optimized every day what the battery was doing.
299
00:20:44.030 --> 00:20:49.385
We did, however, in that project, think you might want an override button. You might not want to
300
00:20:49.765 --> 00:20:52.985
meddle with what exactly the optimization is doing, but
301
00:20:53.500 --> 00:20:59.360
it's nice, of course, in an optimized system to have the ability to say, just stop what you're doing or
302
00:20:59.820 --> 00:21:01.120
the other button says,
303
00:21:01.785 --> 00:21:03.005
charge the car now.
304
00:21:03.545 --> 00:21:05.565
I don't care about the price or whatever.
305
00:21:06.105 --> 00:21:13.340
The idea is that these buttons are almost never used, but they should be there. So I think that end users who live in these
306
00:21:13.880 --> 00:21:17.100
built environments should really not be involved that much,
307
00:21:17.675 --> 00:21:19.055
and it's more the
308
00:21:19.995 --> 00:21:20.495
servicing
309
00:21:21.115 --> 00:21:23.135
around that. So Flex Measures
310
00:21:23.435 --> 00:21:25.935
needs to be hosted, and we need to
311
00:21:26.340 --> 00:21:28.679
check that that data is always flowing
312
00:21:29.700 --> 00:21:32.519
and check with every error that might
313
00:21:32.899 --> 00:21:34.440
come up in in the scheduling.
314
00:21:35.175 --> 00:21:38.555
And you have some companies that you call energy service companies.
315
00:21:39.175 --> 00:21:41.355
So a lot of them, for instance, have these more traditional
316
00:21:41.815 --> 00:21:44.235
energy management systems I talked about earlier.
317
00:21:45.080 --> 00:21:45.580
And
318
00:21:46.120 --> 00:21:48.620
I think that they would be users of Flex Measures.
319
00:21:49.400 --> 00:21:51.420
If we don't do it for them, they would really
320
00:21:52.200 --> 00:21:54.060
build it into their monitoring pipeline,
321
00:21:54.745 --> 00:21:59.645
have a KPI and reporting pipeline to regularly check if that's happening correctly.
322
00:22:00.345 --> 00:22:02.765
1 thing that's nice to check is, the universal
323
00:22:04.540 --> 00:22:05.920
smart energy framework.
324
00:22:06.460 --> 00:22:07.680
It's called Yusef.
325
00:22:08.460 --> 00:22:13.875
My cofounder was on the design committee, and they talked about the different roles in modern smart energy system.
326
00:22:14.415 --> 00:22:16.035
You mentioned the prosumer already.
327
00:22:16.495 --> 00:22:20.915
I mentioned the energy service companies, and you also have companies they're called aggregators.
328
00:22:22.020 --> 00:22:22.520
They
329
00:22:23.059 --> 00:22:24.520
aggregate the energy flexibility
330
00:22:25.620 --> 00:22:26.120
from
331
00:22:26.900 --> 00:22:27.880
a lot of customers
332
00:22:28.180 --> 00:22:31.715
into their virtual power plant and and just trade that flexibility
333
00:22:32.335 --> 00:22:33.155
on markets.
334
00:22:34.015 --> 00:22:37.075
I think you could build energy aggregator using FlexNetters,
335
00:22:38.159 --> 00:22:40.980
but that's not what we do. We use it for
336
00:22:41.600 --> 00:22:43.299
optimization behind your meter.
337
00:22:43.679 --> 00:22:44.980
As far as the actual
338
00:22:45.515 --> 00:22:46.015
implementation
339
00:22:46.635 --> 00:22:50.095
of Flex Measures, I'm wondering if you can talk to some of the
340
00:22:50.875 --> 00:22:55.260
design considerations that went into it, some of the challenges of being able to
341
00:22:55.720 --> 00:23:00.940
flexibly manage scale of inbound data and the types of requests that you're getting.
342
00:23:01.905 --> 00:23:15.460
And also, I'm interested in maybe some of the capabilities of being able to federate across different installations of flex measures and, you know, route different decision requirements or different data points based on different kind of inbound information?
343
00:23:16.365 --> 00:23:17.425
So from the beginning,
344
00:23:18.205 --> 00:23:19.265
a few years ago,
345
00:23:19.565 --> 00:23:21.825
I think, I didn't mention that before. But the first
346
00:23:22.445 --> 00:23:26.150
project where Flex Measures really got its first development
347
00:23:26.850 --> 00:23:30.549
time, that was with a South Korean customer called A1 Engineering.
348
00:23:31.490 --> 00:23:33.110
They wanted to do simulation
349
00:23:34.345 --> 00:23:37.485
first of a of a system with a lot of smart charging,
350
00:23:37.785 --> 00:23:39.245
but then also move to
351
00:23:39.545 --> 00:23:42.559
real time operation in a later phase. So this is
352
00:23:43.019 --> 00:23:44.640
the main design idea
353
00:23:45.100 --> 00:23:47.440
that has been there from the start that you can use
354
00:23:48.020 --> 00:23:48.520
something
355
00:23:49.100 --> 00:23:50.960
in reality in real operation,
356
00:23:51.375 --> 00:23:52.195
but it's also useful
357
00:23:53.375 --> 00:23:56.755
initially in a phase where you are simulating what might actually happen.
358
00:23:57.215 --> 00:23:58.755
And that's still very relevant.
359
00:23:59.310 --> 00:24:07.650
So we're doing a simulation with a bigger real estate company right now. That's their first step. I think that that main design goal still stands.
360
00:24:08.175 --> 00:24:13.315
Now I haven't used multiple Flex Measures nodes and then federated the decisions yet. So I think
361
00:24:13.855 --> 00:24:18.490
that could be built, like, a lower level of Flex Measures instances, basically, governing,
362
00:24:20.309 --> 00:24:21.370
smaller installations.
363
00:24:22.044 --> 00:24:23.105
And another 1,
364
00:24:23.645 --> 00:24:28.544
maybe trying to aggregate. As I mentioned earlier, the it's a concept of the flexibility aggregator
365
00:24:29.005 --> 00:24:30.225
and aggregate across
366
00:24:31.080 --> 00:24:35.419
those behind the meter installations that could be a higher level of extensions
367
00:24:36.200 --> 00:24:38.860
instance. Yeah. Now we haven't done that yet.
368
00:24:39.385 --> 00:24:40.285
We consciously
369
00:24:40.905 --> 00:24:45.405
can decide not to, you know, let the volume of data grow too large.
370
00:24:46.265 --> 00:24:47.085
So for instance,
371
00:24:47.529 --> 00:24:50.029
you might run into projects where they have 10 second
372
00:24:50.649 --> 00:24:51.630
interval readings.
373
00:24:52.409 --> 00:24:59.475
But to make the kind of decisions that Flex Measures needs to make, let's say, on an hourly basis or 15 minute basis,
374
00:25:00.335 --> 00:25:02.755
you don't need to read in the finely
375
00:25:03.215 --> 00:25:08.350
structured data. It would be best if that other system gives us already aggregated
376
00:25:09.130 --> 00:25:15.865
data. For instance, in in 5 minute steps, that'll be fine. And then the amount of data isn't growing that much,
377
00:25:16.805 --> 00:25:19.125
but we're already looking into that. So we
378
00:25:19.765 --> 00:25:21.465
our data is stored in, Postgres.
379
00:25:22.470 --> 00:25:26.810
Of course, that we also made that choice because you have so much support for
380
00:25:27.510 --> 00:25:32.565
growing bigger with your databases and have all these clusters, clustering strategies, and sharding.
381
00:25:33.425 --> 00:25:35.125
And, also, of course, there's interesting
382
00:25:36.225 --> 00:25:42.210
technology built upon Postgres. I think 1 of called Timescale DB that really focuses only on the time series
383
00:25:43.310 --> 00:25:43.810
aspect.
384
00:25:44.510 --> 00:25:46.290
And 1 thing that we considerably
385
00:25:47.175 --> 00:25:49.435
put some effort in is to model
386
00:25:50.295 --> 00:25:53.195
this data because it comes, as you said, from different sources.
387
00:25:54.055 --> 00:25:55.835
There might be a company that
388
00:25:56.159 --> 00:25:57.299
installed the sensors
389
00:25:58.000 --> 00:26:00.340
at the factory or is responsible
390
00:26:00.640 --> 00:26:02.980
at the houses to measure what's happening.
391
00:26:03.455 --> 00:26:04.674
So weather forecasting
392
00:26:05.215 --> 00:26:07.715
service. So we use, open weather map.
393
00:26:08.335 --> 00:26:10.275
Flex measures can read data from there.
394
00:26:10.820 --> 00:26:14.120
The European transmission system operators have a service
395
00:26:14.500 --> 00:26:18.360
in open transparent API as well. So you got multiple sources.
396
00:26:18.925 --> 00:26:21.105
They all send you data that might be
397
00:26:21.645 --> 00:26:22.785
about different things.
398
00:26:23.325 --> 00:26:23.985
It has
399
00:26:24.525 --> 00:26:25.745
different time zones.
400
00:26:26.470 --> 00:26:29.370
It's crucial to maybe know when they know about that.
401
00:26:29.830 --> 00:26:32.250
So not only the time that they're talking about,
402
00:26:32.950 --> 00:26:34.890
1 point in time with 1 measurement,
403
00:26:36.865 --> 00:26:37.765
has 1
404
00:26:38.545 --> 00:26:45.684
date time, of course, about the time that you're talking about. But we're also interested in the daytime when they said this would happen.
405
00:26:46.410 --> 00:26:50.510
If it's before the time, this is a forecast. If it's after the time, it's a measurement
406
00:26:51.289 --> 00:26:54.830
that can be quite useful when you write optimization algorithms.
407
00:26:55.735 --> 00:26:58.075
All this led us to write our own,
408
00:26:58.615 --> 00:26:59.115
basically,
409
00:26:59.575 --> 00:27:03.035
time series data model. We called it Timely Beliefs.
410
00:27:03.649 --> 00:27:05.269
It's actually an extension
411
00:27:05.570 --> 00:27:09.990
on pandas. So what it makes possible is that you have a pandas data frame
412
00:27:10.664 --> 00:27:15.804
or basically a class that extends from Pandas data frame, and we call it the beliefs data frame.
413
00:27:16.105 --> 00:27:24.380
We wrote a lot of functionality around that that you can, you know, merge these data frames or change the resolution or whatever. It's a very interesting project.
414
00:27:24.840 --> 00:27:29.565
They took a lot of our time, and now we're profiting from that within Flex Measures to really understand all the time
415
00:27:30.264 --> 00:27:31.804
what that data really means.
416
00:27:32.264 --> 00:27:39.110
As far as the kind of architecture of the project, digging a bit more into that, I'm also interested in understanding the
417
00:27:39.410 --> 00:27:42.790
kind of extension points that you've built to be able to allow
418
00:27:43.215 --> 00:27:47.795
different users of Flex Measures to adapt to different data sources or
419
00:27:48.175 --> 00:27:56.510
being able to extend the types of decisions and machine learning capabilities that they want to use or the types of predictions that they want to generate
420
00:27:56.970 --> 00:28:04.335
and maybe a bit more of the kind of data integration and kind of predictive capabilities that you've built into it? Flex Measures is extensible
421
00:28:05.115 --> 00:28:05.855
with plugins
422
00:28:06.395 --> 00:28:08.495
done last year, and we are using that ourselves
423
00:28:09.210 --> 00:28:10.350
because often
424
00:28:11.050 --> 00:28:12.830
every project has these differences,
425
00:28:13.210 --> 00:28:15.390
different data sources, or or needs.
426
00:28:16.010 --> 00:28:17.630
Often, it's also the visualization
427
00:28:18.685 --> 00:28:20.545
that you want to adapt to a customer.
428
00:28:21.165 --> 00:28:22.225
That's very important
429
00:28:22.845 --> 00:28:25.425
to the whole architecture that you can add your own
430
00:28:26.100 --> 00:28:27.400
API endpoints even,
431
00:28:27.860 --> 00:28:31.160
your own visualization, and and we have a command line interface.
432
00:28:31.860 --> 00:28:33.480
You can even add your own
433
00:28:33.975 --> 00:28:34.475
command
434
00:28:35.014 --> 00:28:36.235
line interface commands.
435
00:28:36.934 --> 00:28:38.794
For instance, 1 that regularly
436
00:28:39.174 --> 00:28:41.514
gets data from some other open API,
437
00:28:41.960 --> 00:28:46.540
and you want to run that hourly in a Chrome job as a Flex Measures command.
438
00:28:47.000 --> 00:28:49.500
Flex Measures actually is built on Flask,
439
00:28:50.395 --> 00:28:53.615
and Flask has some nice capabilities for that.
440
00:28:53.995 --> 00:29:00.420
Of course, for APIs, but also the it integrates with click. So command line interfaces are, nice to write.
441
00:29:00.960 --> 00:29:02.820
And with the Flask blueprint
442
00:29:04.080 --> 00:29:05.380
technology, it was
443
00:29:05.985 --> 00:29:08.405
quite a joy to write our 1st plug in
444
00:29:08.865 --> 00:29:11.925
architecture. So what you do in the plug in is mostly
445
00:29:12.545 --> 00:29:13.365
add a blueprint
446
00:29:14.190 --> 00:29:15.090
to flex measures.
447
00:29:15.630 --> 00:29:16.850
Flask has a CLI
448
00:29:17.150 --> 00:29:19.410
command, and we renamed that to flex measures.
449
00:29:20.030 --> 00:29:20.530
So,
450
00:29:21.205 --> 00:29:27.705
yeah, it all ties back into what I said earlier. We want flex measures to be usable by software developers. And then,
451
00:29:28.300 --> 00:29:35.360
of course, APIs are super important, but also the command line interface, I think, is is super important. We wrote a
452
00:29:36.085 --> 00:29:40.184
tutorial how you would use Flex Measures within a few minutes, you know, just to show you
453
00:29:40.565 --> 00:29:42.905
within a few minutes what you might be doing,
454
00:29:43.620 --> 00:29:50.840
which is based on Docker. So you don't need to install all these things. Just pull the Docker image. And then that tutorial is
455
00:29:51.415 --> 00:29:52.075
a few
456
00:29:52.455 --> 00:29:53.995
calls to the command line interface
457
00:29:54.855 --> 00:29:59.915
about modeling that couple assets that you have a battery, for instance, loading in some data
458
00:30:00.590 --> 00:30:04.690
from a CSV file and asking Flex Measures to compute a great
459
00:30:04.990 --> 00:30:09.890
optimized schedule for the next day. That's all command line interface command. That's what I really like.
460
00:30:10.465 --> 00:30:10.965
Optimization
461
00:30:11.505 --> 00:30:12.325
in the background
462
00:30:13.185 --> 00:30:13.665
happens
463
00:30:14.145 --> 00:30:15.445
I think we use pyomo.
464
00:30:16.465 --> 00:30:17.925
It's also a very powerful
465
00:30:18.889 --> 00:30:20.590
library for for these kind of things.
466
00:30:21.289 --> 00:30:23.309
And we used there was some sklearn
467
00:30:23.690 --> 00:30:25.230
that we used. But recently,
468
00:30:25.929 --> 00:30:27.470
we've picked up sktime,
469
00:30:28.434 --> 00:30:29.335
which is really
470
00:30:29.794 --> 00:30:32.054
a package focusing on time series
471
00:30:32.595 --> 00:30:36.534
forecasting, doing smart things around that, and it's very nicely done, it seems.
472
00:30:37.190 --> 00:30:38.890
As you have been
473
00:30:39.429 --> 00:30:41.210
building Flex Measures and
474
00:30:41.510 --> 00:30:44.490
iterating on the problem space and also with the
475
00:30:45.184 --> 00:30:47.684
donation of it to the LF Energy
476
00:30:48.304 --> 00:30:48.804
Foundation.
477
00:30:49.345 --> 00:30:57.299
I'm wondering what are some of the ways that the design and goals and scope of the project have changed or evolved since you first started working on it? I said earlier that,
478
00:30:58.000 --> 00:30:58.500
actually,
479
00:30:59.440 --> 00:31:07.645
the overall design goal to make both simulation and optimized real time operation possible from the same code base is still intact.
480
00:31:08.585 --> 00:31:12.360
I liked open source in Flex Measures. It it happened roughly,
481
00:31:12.820 --> 00:31:15.880
well, less than 2 years ago. I we hit that button in GitHub
482
00:31:16.755 --> 00:31:19.895
in a couple months later, donated to the Linux Energy Foundation.
483
00:31:20.835 --> 00:31:23.815
I liked that that pushed us to even more transparency
484
00:31:24.195 --> 00:31:27.030
and even more having a focus on code quality
485
00:31:27.650 --> 00:31:29.030
to make sure that
486
00:31:29.890 --> 00:31:30.710
Blake 8,
487
00:31:31.570 --> 00:31:32.070
MyPy,
488
00:31:33.145 --> 00:31:36.205
and Black always run. It's always looking good.
489
00:31:36.585 --> 00:31:37.245
And documentation
490
00:31:38.025 --> 00:31:39.325
got a lot of boost.
491
00:31:40.505 --> 00:31:41.165
And, also,
492
00:31:41.900 --> 00:31:47.600
in the meantime, we did once make big changes to the data model. When we worked for industry customers,
493
00:31:48.380 --> 00:31:51.280
we noticed that our initial data model was a bit too simplistic.
494
00:31:51.645 --> 00:31:52.625
I think we modeled
495
00:31:53.645 --> 00:31:54.145
power,
496
00:31:55.005 --> 00:31:56.545
weather data, and prices.
497
00:31:57.565 --> 00:32:03.070
And if you go to industry, you quickly realize they want to measure a lot of other things that all have
498
00:32:03.690 --> 00:32:05.390
some influence on your
499
00:32:05.770 --> 00:32:06.270
outcomes.
500
00:32:06.735 --> 00:32:09.235
For instance, you might be measuring steam flow,
501
00:32:10.095 --> 00:32:10.595
and
502
00:32:10.975 --> 00:32:32.395
that wasn't possible. So we took a lot of time to make our data model more generic. Now we can model anything. Another interesting aspect of the kind of open sourcing question, and it's addressed a little bit by your donation of the project to the LF Energy Foundation. But I also know that you are running a business where Flex Measures serves as a kind of central portion of what you're providing
503
00:32:32.850 --> 00:32:36.470
at CITA. And I'm curious if you can talk to some of the ways that you think about
504
00:32:36.850 --> 00:32:42.755
which pieces of your work you want to put into the open source Flex Measures project and what are some of
505
00:32:43.315 --> 00:32:50.775
the other kind of capabilities or layers that you're building on top of that that are part of the kind of commercial and business focused offering?
506
00:32:51.190 --> 00:32:54.410
It's a very interesting time to have a commercial open source
507
00:32:54.790 --> 00:32:55.530
start up.
508
00:32:55.990 --> 00:32:57.930
I was recently involved in an accelerator,
509
00:32:58.390 --> 00:32:59.450
on the US side.
510
00:33:00.135 --> 00:33:03.035
I had to attend at night, always at 11 PM.
511
00:33:03.655 --> 00:33:04.155
But
512
00:33:04.695 --> 00:33:06.875
there's really some investors in the US
513
00:33:08.029 --> 00:33:11.330
that are looking for open source companies to invest in
514
00:33:11.710 --> 00:33:15.409
because you got, in the last couple of years, some really interesting examples,
515
00:33:15.805 --> 00:33:17.025
mostly on the database
516
00:33:17.725 --> 00:33:19.105
developer tooling side.
517
00:33:19.885 --> 00:33:24.700
So, obviously, in the energy sector, this is quite new. But it's clear that you can
518
00:33:25.240 --> 00:33:27.100
run a business even if you're
519
00:33:27.720 --> 00:33:28.620
very open.
520
00:33:29.240 --> 00:33:35.125
LIMX Energy Foundation advised to use permissive licenses. So in this case, we use Apache 2.
521
00:33:35.745 --> 00:33:37.685
You could also use MIT, for instance,
522
00:33:38.200 --> 00:33:43.820
over a Copyleft licenses because the idea is to just, you know, open the doors wide to
523
00:33:44.440 --> 00:33:45.340
growing the community,
524
00:33:46.005 --> 00:33:53.865
letting other people use the tooling you've you write. Even if they don't re even tell you about it, they should have no doubts that they can use it freely
525
00:33:54.270 --> 00:34:00.929
because you want to just grow, grow, grow, grow. That's after the Linx Energy Foundation told us that they use those licenses.
526
00:34:01.665 --> 00:34:06.645
I also heard it from this venture capitalist accelerator program. It's a very interesting
527
00:34:07.345 --> 00:34:14.740
vision on innovation, on open innovation that this might be 1 of the future scenarios that comes true. Right? That the pool of open source
528
00:34:16.000 --> 00:34:16.980
innovation becomes
529
00:34:17.440 --> 00:34:20.339
as big in other sectors has become in databases
530
00:34:20.815 --> 00:34:21.475
and developer
531
00:34:22.015 --> 00:34:25.875
tooling. Nonetheless, well, we have made Flex Measures possible to create plug ins.
532
00:34:26.415 --> 00:34:26.915
And
533
00:34:27.455 --> 00:34:41.105
oftentimes, the customer says, well, if you're developing something here about our processes, I don't want that to be public. Or, if I pay you to do something, I don't want it to be public. And with the plug ins, that's very possible, of course. Then you have
534
00:34:41.405 --> 00:34:43.105
Flex Measures as an open platform.
535
00:34:43.485 --> 00:34:45.745
And, also, Timely Believes is open,
536
00:34:46.060 --> 00:34:47.119
and we occasionally
537
00:34:47.420 --> 00:34:53.680
open up other projects that we work on. For instance, I mentioned that, Home Assistant plug in in the in the vehicles to grid
538
00:34:53.980 --> 00:34:55.200
project that's open.
539
00:34:55.505 --> 00:34:56.805
I think we have a couple
540
00:34:57.265 --> 00:34:58.085
closed repositories
541
00:34:58.385 --> 00:35:00.485
lying around right now with work for customers.
542
00:35:01.105 --> 00:35:03.125
It's something that's maybe not ready for the world.
543
00:35:03.425 --> 00:35:04.485
It will always happen.
544
00:35:05.000 --> 00:35:11.580
I'm also interested in understanding a bit about given the kind of predictive and machine learning focused capabilities
545
00:35:12.025 --> 00:35:14.525
that are part of the Flex Measures platform
546
00:35:15.225 --> 00:35:16.685
and the fact that you are
547
00:35:17.225 --> 00:35:17.725
building
548
00:35:18.105 --> 00:35:19.645
predictions about and
549
00:35:23.610 --> 00:35:24.510
energy capacity,
550
00:35:25.290 --> 00:35:26.830
what are some of the
551
00:35:27.505 --> 00:35:39.210
challenges that you face in terms of accounting for whether to bias towards false positives or false negatives, and what levels of uncertainty you're willing to accept in some of those predictions and decisions that you
552
00:35:39.510 --> 00:35:43.050
make? Yeah. That is a very interesting topic to go into.
553
00:35:43.535 --> 00:35:44.995
It's difficult to answer
554
00:35:45.535 --> 00:35:48.675
without having even more experience. We had
555
00:35:49.295 --> 00:35:51.555
a project trying to find out if
556
00:35:52.270 --> 00:36:00.050
forecasting should be done with something very sophisticated. You know? You can use linear regression to forecast, let's say, the solar production,
557
00:36:00.590 --> 00:36:01.490
RIMO models,
558
00:36:01.790 --> 00:36:04.105
but But you can also use a neural network.
559
00:36:04.645 --> 00:36:07.145
We found that the the neural network isn't
560
00:36:07.765 --> 00:36:08.245
really better
561
00:36:08.965 --> 00:36:12.619
that much better, and you don't know what's going on. And just recently,
562
00:36:13.320 --> 00:36:14.780
I saw a paper come by
563
00:36:15.320 --> 00:36:19.420
that some researchers even confirmed that. So for time series forecasting,
564
00:36:20.445 --> 00:36:24.225
currently, there doesn't seem to be a reason to become very sophisticated.
565
00:36:25.165 --> 00:36:26.865
What we did do was
566
00:36:27.329 --> 00:36:28.150
we saw that
567
00:36:28.849 --> 00:36:29.910
there is algorithms
568
00:36:30.849 --> 00:36:32.549
which tell you about the uncertainty
569
00:36:33.650 --> 00:36:35.730
that lies in the prediction. So it's mostly,
570
00:36:36.130 --> 00:36:36.630
ensemble
571
00:36:37.565 --> 00:36:39.664
algorithms like random forest.
572
00:36:40.525 --> 00:36:47.240
They have, you know, a lot of forecast and then make a a mean forecast, but they tell you how far all of the others
573
00:36:47.620 --> 00:36:48.600
were distributed.
574
00:36:49.060 --> 00:36:49.560
And
575
00:36:50.340 --> 00:36:57.695
that information can potentially, of course, be useful for exactly your question that you know that some forecasts are
576
00:36:58.555 --> 00:36:59.055
shaky,
577
00:36:59.755 --> 00:37:01.375
and you might decide
578
00:37:02.369 --> 00:37:06.950
for a scheduling algorithm that there's a policy of not acting in those situations.
579
00:37:07.809 --> 00:37:14.644
And that's something I didn't mention about the timely beliefs data model. 1 column in the belief data frame is the uncertainty. So how certain
580
00:37:15.345 --> 00:37:21.089
forecast, for instance, would be. And it reminds me of a setting we have in a scheduling algorithm right now.
581
00:37:21.470 --> 00:37:22.769
For the battery charging,
582
00:37:23.710 --> 00:37:26.849
1 property you need then is, round trip efficiency.
583
00:37:27.695 --> 00:37:28.195
So
584
00:37:28.895 --> 00:37:29.875
you can discharge
585
00:37:30.175 --> 00:37:30.915
a battery.
586
00:37:31.455 --> 00:37:33.635
Let's say, you discharge 1 kilowatt hour
587
00:37:34.255 --> 00:37:35.875
and then you charge that back.
588
00:37:36.280 --> 00:37:38.060
Let's say the other 1 around. You charge it
589
00:37:38.360 --> 00:37:40.060
first and then you discharge it,
590
00:37:40.360 --> 00:37:45.740
and you've made some arbitrage because your 24 hourly prices had this nice difference
591
00:37:46.175 --> 00:37:49.315
or that some difference. But the battery is not 100% efficient,
592
00:37:49.695 --> 00:37:52.355
so you end up maybe that it's 80% efficient,
593
00:37:52.975 --> 00:37:54.915
could become 90% in some batteries.
594
00:37:55.580 --> 00:37:58.160
So you're losing some of that power in that process.
595
00:37:58.940 --> 00:37:59.440
And
596
00:37:59.900 --> 00:38:04.400
it's very important to model that so that the scheduling algorithm then sometimes does nothing.
597
00:38:04.725 --> 00:38:07.065
Price difference difference is too low,
598
00:38:07.525 --> 00:38:14.105
then it doesn't make sense to do all that. So sometimes the schedule is to do nothing even if there are some price differences. And I see the
599
00:38:14.579 --> 00:38:19.960
policy of doing nothing if the uncertainty is too high, and that's a similar thing.
600
00:38:20.339 --> 00:38:22.795
In your work of building flex measures
601
00:38:23.095 --> 00:38:27.435
and working with different companies to help implement it and
602
00:38:27.735 --> 00:38:36.050
providing it to the community and working with people who are using it for their own purposes. What are some of the most interesting or innovative or unexpected ways that you've seen flex measures applied?
603
00:38:36.434 --> 00:38:37.654
There's not that much,
604
00:38:38.115 --> 00:38:42.855
but there was a project where flex measures wasn't even being used for any scheduling,
605
00:38:43.395 --> 00:38:44.214
which is
606
00:38:45.119 --> 00:38:49.299
basically the end goal of using Flex Measures is to have a schedule and
607
00:38:49.760 --> 00:38:53.059
then change the times when you do use energy.
608
00:38:53.674 --> 00:38:56.414
And in that case, it was a project where it was
609
00:38:56.714 --> 00:38:58.095
just so useful to
610
00:38:58.635 --> 00:39:02.174
use Flex Measures to just be a plain energy management system,
611
00:39:02.530 --> 00:39:06.710
getting all this data together with data pipelining being done quite fast,
612
00:39:07.329 --> 00:39:13.375
but then not doing forecasting but in doing something else. There was a different algorithm that was basically some advanced analytics.
613
00:39:14.155 --> 00:39:18.575
And it didn't you have to use electricity. It was about gas consumption in in industry.
614
00:39:19.069 --> 00:39:21.809
So that is possible, but that's not where I want to take it.
615
00:39:22.270 --> 00:39:32.115
In your work of building the project and building a business around it and donating it to the community, what are some of the most interesting or unexpected or challenging lessons that you've learned in the process?
616
00:39:32.815 --> 00:39:37.155
Something that you have to learn on your own is how long it takes to integrate
617
00:39:37.695 --> 00:39:38.195
data
618
00:39:38.810 --> 00:39:41.310
that constantly flows and that has these
619
00:39:42.090 --> 00:39:45.070
intricacies that I mentioned earlier about time series
620
00:39:45.690 --> 00:39:48.350
that you could get slightly wrong. It could be
621
00:39:48.735 --> 00:39:51.075
you get the time zone slightly wrong or
622
00:39:51.375 --> 00:39:51.875
somebody
623
00:39:52.255 --> 00:39:54.515
sends you data that isn't localized.
624
00:39:55.670 --> 00:39:56.410
You might
625
00:39:57.109 --> 00:39:59.770
be 1 hour off half of the year.
626
00:40:00.230 --> 00:40:02.010
Your whole schedule will not be good.
627
00:40:02.470 --> 00:40:13.645
They also might, send you data that gets the daylight saving time wrong, and then there might be 2 hours in the year where you fail when daylight saving time is switching back and forth.
628
00:40:14.320 --> 00:40:16.100
And, yeah, the sheer amount
629
00:40:16.560 --> 00:40:17.060
of
630
00:40:17.440 --> 00:40:20.020
attention and work that gets into that
631
00:40:20.560 --> 00:40:26.005
is is quite staggering. We had an industry client with a partner who does the censoring,
632
00:40:26.865 --> 00:40:29.685
and they didn't have the API ready for us. So
633
00:40:30.145 --> 00:40:30.724
it was
634
00:40:31.160 --> 00:40:38.220
not what I had expected in 2021 from that kind of player. What we could get from them was emails with Excel spreadsheets,
635
00:40:39.145 --> 00:40:49.920
And we decided that they send us 3 per day. So every 8 hours, we have we wrote a Flexner's plug in that has its own command line interface command to just handle that email
636
00:40:50.460 --> 00:40:51.839
and get the data in.
637
00:40:52.539 --> 00:40:55.760
Yeah. You can imagine that that prolongs that kind of pilot
638
00:40:56.299 --> 00:40:56.799
considerably.
639
00:40:57.444 --> 00:40:59.545
Yeah. It's definitely remarkable
640
00:40:59.845 --> 00:41:00.505
how much
641
00:41:00.885 --> 00:41:01.545
of modern
642
00:41:01.924 --> 00:41:06.184
data engineering and data management still relies on CSVs and FTP.
643
00:41:06.980 --> 00:41:07.480
Yes.
644
00:41:08.260 --> 00:41:09.320
Absolutely. And
645
00:41:09.620 --> 00:41:12.840
if you look at the energy sector as a whole that these kind of experiences,
646
00:41:13.700 --> 00:41:14.520
1 dimension
647
00:41:15.865 --> 00:41:27.250
towards the realization that there will not be 1 end to end solution that can be used across the whole country or across a continent that that will just take over. We have to
648
00:41:27.710 --> 00:41:28.850
build a lot of
649
00:41:29.550 --> 00:41:31.010
unique end solutions.
650
00:41:31.924 --> 00:41:33.224
Let's call them apps
651
00:41:33.684 --> 00:41:37.385
that work differently in different energy systems, countries,
652
00:41:37.924 --> 00:41:39.144
cultural contexts,
653
00:41:40.160 --> 00:41:40.660
sectors.
654
00:41:41.440 --> 00:41:42.420
And I think
655
00:41:43.040 --> 00:41:46.260
flex measures can be a basis for all of that, but you always need to
656
00:41:46.720 --> 00:41:48.660
put something on top that really fits
657
00:41:49.325 --> 00:41:50.225
to that situation.
658
00:41:50.845 --> 00:41:53.745
For people who are interested in being able
659
00:41:54.285 --> 00:42:05.420
to gain some more insight and potentially act on the different aspects of energy usage, whether it's at the residential or industrial or grid scale? What are the cases where Flex Measures is the wrong choice?
660
00:42:05.835 --> 00:42:10.255
I've mentioned that earlier, but there are some people who have to act within seconds.
661
00:42:10.875 --> 00:42:14.460
And that's usually if you're really taking part in this first
662
00:42:15.160 --> 00:42:15.980
reserve mechanism
663
00:42:16.440 --> 00:42:18.140
reacting to balancing problems,
664
00:42:19.240 --> 00:42:20.460
frequency things.
665
00:42:21.275 --> 00:42:21.775
And
666
00:42:22.315 --> 00:42:30.175
we have put a lot of focus on flex measures to be developer friendly and correct about the data so that developers really understand what they're building.
667
00:42:31.210 --> 00:42:33.869
And we haven't made it, you know, a Rust project
668
00:42:34.569 --> 00:42:36.670
that lets you do all that within seconds.
669
00:42:37.210 --> 00:42:40.269
As you continue to build and iterate on the project
670
00:42:40.734 --> 00:42:50.035
and continue to iterate on the business that you're building on top of it, what are some of the things you have planned for the near to medium term or any particular problem areas that you're excited to dig into?
671
00:42:50.580 --> 00:42:53.400
Yeah. So first, about the FlexVicious project,
672
00:42:54.340 --> 00:42:54.920
I think
673
00:42:55.380 --> 00:42:58.520
what we really could do with that project is to support
674
00:42:59.065 --> 00:43:01.805
the 2 groups of users that I see
675
00:43:02.585 --> 00:43:03.725
in what they are
676
00:43:04.025 --> 00:43:08.525
doing as a as a first steps. So I see the first group as being the 90%
677
00:43:09.450 --> 00:43:09.950
users
678
00:43:10.650 --> 00:43:12.430
who have a use case that
679
00:43:13.530 --> 00:43:14.270
really happens
680
00:43:15.290 --> 00:43:19.585
a lot in the world, and Flex Measures should support out of the box
681
00:43:20.204 --> 00:43:21.105
that the optimization
682
00:43:21.484 --> 00:43:25.345
for that use case can be done within, you know, a few hours.
683
00:43:25.680 --> 00:43:31.220
Data integration plus having that optimization work should be a few hours. So that that's a a little bit of a benchmark.
684
00:43:31.840 --> 00:43:33.140
Let's say, as an example,
685
00:43:34.085 --> 00:43:37.704
somebody has solar on the roof and there's somehow
686
00:43:38.244 --> 00:43:41.065
a battery, a home battery, or an electric car,
687
00:43:41.950 --> 00:43:43.730
and they have a dynamic tariff.
688
00:43:44.589 --> 00:43:48.530
That might happen a lot in the next years across the world, something like that.
689
00:43:49.685 --> 00:43:51.465
And that should just work. And,
690
00:43:52.565 --> 00:43:55.465
basically, that does work now. What we have to do is documentation,
691
00:43:56.165 --> 00:43:57.545
write an excellent tutorial
692
00:43:58.510 --> 00:43:59.250
so that
693
00:43:59.630 --> 00:44:01.010
everybody could do that.
694
00:44:01.470 --> 00:44:04.850
And there's probably a list of 4 or 5 more of these
695
00:44:05.710 --> 00:44:06.210
situations
696
00:44:07.265 --> 00:44:08.405
that should just work.
697
00:44:08.705 --> 00:44:11.605
And the other group of users are the power users
698
00:44:12.625 --> 00:44:16.565
that really have a unique problem and may come with their unique
699
00:44:17.490 --> 00:44:18.870
idea of solving it.
700
00:44:19.250 --> 00:44:21.510
And then I'm talking mostly about the algorithms.
701
00:44:22.610 --> 00:44:23.990
Right now, you can already
702
00:44:24.785 --> 00:44:26.565
plug in your own scheduling algorithm.
703
00:44:27.265 --> 00:44:30.485
Or to be honest, at the time of recording, that is a PR I'm
704
00:44:31.345 --> 00:44:32.165
having reviewed.
705
00:44:32.869 --> 00:44:35.930
But that's really important for some people. That's researchers,
706
00:44:36.790 --> 00:44:37.290
certainly.
707
00:44:38.470 --> 00:44:42.010
That's their whole line of work to come up with better algorithms.
708
00:44:42.855 --> 00:44:45.915
Maybe find that neural network algorithm that really does make the difference.
709
00:44:46.535 --> 00:44:49.515
But you see that also in some companies really have
710
00:44:49.950 --> 00:44:54.609
certainly an industry that just have their unique challenge. And they should be using flex measures,
711
00:44:55.390 --> 00:44:59.365
you know, as is and then just change that 1 thing that there's
712
00:44:59.825 --> 00:45:00.885
a common interface
713
00:45:01.345 --> 00:45:02.244
with how
714
00:45:03.025 --> 00:45:05.765
the question to make a new schedule is described.
715
00:45:06.640 --> 00:45:11.780
But they could, in the back, just switch out the algorithm that is used to make the schedule.
716
00:45:12.240 --> 00:45:14.580
I think these 2 groups of users deserve
717
00:45:14.975 --> 00:45:16.355
each of their own dedicated
718
00:45:17.135 --> 00:45:24.435
tutorials by us to really make it clear, to make it a joy to do so much as our work. I mentioned 2 sectors that we're working on.
719
00:45:24.859 --> 00:45:25.599
The industry
720
00:45:26.619 --> 00:45:28.720
world is very interesting and challenging,
721
00:45:29.740 --> 00:45:31.280
but it's also very diverse.
722
00:45:31.820 --> 00:45:33.119
Right now, those environments,
723
00:45:33.445 --> 00:45:35.785
what excites me the most, and I'm I'm talking to
724
00:45:36.325 --> 00:45:39.065
companies there that really understood the scale
725
00:45:39.525 --> 00:45:43.800
that I mentioned earlier, they have renovate or we modernized the whole portfolio
726
00:45:44.180 --> 00:45:48.119
within a very short amount of years. So scale is their big topic.
727
00:45:48.660 --> 00:45:51.160
And the third thing I didn't mention is microgrids.
728
00:45:51.925 --> 00:45:55.145
So microgrid is your own little electricity grid
729
00:45:55.445 --> 00:45:56.345
that is connected
730
00:45:56.965 --> 00:45:58.745
maybe to the main grid
731
00:45:59.045 --> 00:45:59.785
or not,
732
00:46:00.150 --> 00:46:02.890
completely run on its own. And that's a situation
733
00:46:03.830 --> 00:46:06.730
that we might see all across the world. So there might be
734
00:46:07.110 --> 00:46:07.610
more
735
00:46:08.635 --> 00:46:10.495
self serving neighborhoods
736
00:46:11.755 --> 00:46:13.135
that are not completely
737
00:46:13.835 --> 00:46:21.520
dependent on the grid in Western countries, but certainly in Africa, that's something that's just the 1 way to go for a lot of regions.
738
00:46:22.060 --> 00:46:28.205
There are some interesting companies there that I hope to do more with in the near future because Flex Measures should be able to
739
00:46:29.245 --> 00:46:29.745
basically
740
00:46:30.045 --> 00:46:31.185
run a small village.
741
00:46:31.725 --> 00:46:36.445
You might think there's too many assets, but what really matters is the flexible assets. So that's where
742
00:46:37.400 --> 00:46:42.700
that defines how much flex flex measures has to do, how many flexible things are there.
743
00:46:43.320 --> 00:46:43.820
Emobility
744
00:46:44.200 --> 00:46:46.940
is making huge steps in Africa, certainly.
745
00:46:47.835 --> 00:46:48.335
We're
746
00:46:48.795 --> 00:46:55.295
hoping to be involved in a project in Victoria Lake where all the fishers get upgraded to electric boats, for instance,
747
00:46:55.870 --> 00:46:58.450
And they use energy for productive uses,
748
00:46:59.390 --> 00:47:01.570
like cooling the fish they caught,
749
00:47:01.950 --> 00:47:02.850
water purification.
750
00:47:04.110 --> 00:47:12.605
That's all flex it's all something you can shift around in the day. And then flexiators can really help, for instance, in the design phase already to not buy oversized
751
00:47:13.065 --> 00:47:13.565
equipment.
752
00:47:14.180 --> 00:47:28.505
I don't think you will become rich in Africa, but that's something I'd really like to be involved in. Well, for anybody who wants to get in touch with you and follow along with the work that you're doing, I'll have you add your preferred contact information to the show notes. And so with that, I'll move us into the picks.
753
00:47:28.805 --> 00:47:44.975
This week, I'm going to choose the show She Hulk from the Disney plus. Just a fun and goofy show. Some interesting storyline there, so I've been enjoying watching that. So I recommend that for people who are looking for something to do in the evening. And with that, I'll pass it you, Nicholas. Do you have any pics this week?
754
00:47:45.595 --> 00:47:48.815
Yeah. I'm watching a very nice German series on Netflix,
755
00:47:49.595 --> 00:47:50.735
and it's called KLEO,
756
00:47:51.275 --> 00:47:52.895
k l e o.
757
00:47:53.480 --> 00:47:55.180
It has a bit of kill bill
758
00:47:55.800 --> 00:47:57.580
visuals because it's an ex
759
00:47:58.360 --> 00:48:00.380
agent from the East German
760
00:48:00.885 --> 00:48:04.425
Ministry of Defense who got thrown into jail. And she
761
00:48:04.725 --> 00:48:10.425
comes out of jail right when the wall has been torn down and tries to find out what happened, who did it.
762
00:48:10.990 --> 00:48:13.890
It's not just Kill Bill, but it's also charming and funny.
763
00:48:14.270 --> 00:48:32.390
Thank you very much for taking the time today to join me and share the work that you're doing on flex measures. It's definitely a very interesting project in a very important area of effort. So I appreciate all the time and energy that you and the other contributors are putting into that. And I hope you enjoy the rest of your day. Thanks, Tobias. It was a joy to be here.
764
00:48:34.690 --> 00:48:36.595
Thank you for listening. Don't forget to
765
00:48:46.660 --> 00:48:47.720
Visit the site at pythonpodcast.com
766
00:48:48.820 --> 00:48:56.965
to subscribe to the show, sign up for the mailing list, and read the show notes. And if you learned something or tried out a project from the show, then tell us about it. Email host hostpythonpodcast.com
767
00:48:58.065 --> 00:48:58.885
with your story.
768
00:48:59.265 --> 00:49:04.165
And to help other people find the show, please leave a review on Apple Podcasts and tell your friends and coworkers.