Okay.
So, hello everyone.
So, I'm Jean-Vatiste and Geoffroy is here.
So, we are at the software development department from RT.
And RT, was it RT?
So, we will give some elements of context.
So, RT is the French TSO, so transmission system operator.
We handle from 20 kV to 400 kV, so it's the high voltage.
And we must provide electricity 24-7 for all the costumers
and all the inhabitants in France and of course in Europe
because we have to cooperate.
And the particularity is that we are asset owner of the grid,
which means that we are responsible to invest and make sure
that the equipment will be okay to work to complete our mission as a TSO.
And we are also responsible to adapt the structure
to make sure that we will ease the transition energy.
So, we need some interconnections and we also try to adapt the grid
to connect for example offshore wind generators.
So, we have many, many challenges in the fast changing world.
So, we have of course new energy mix with big goal around neutral carbon neutrality.
Sorry, for 2050.
So, it's a big challenge.
And we have also some codes, some regulation that make drastic change
and we must adapt to that.
So, it's a more package where we have a lot of work to do in Europe.
And for that, I will read the sentence because it's very important.
Today...
Okay, so...
Today...
So...
Oh, it's okay.
So, today's need is what is very important to understand is that today's need
is not to build a tool that answers present needs,
but to build a tool that is capable to integrate quickly and efficiently tomorrow's needs.
And if you have the idea of the tender to create...
the way to create new tools, sometimes you make specification,
then you made a tender process, then you ask a vendor to develop,
and this cycle is maybe like four years.
The problem is that we don't know what we will ask to do in five years
because everything is changing very fast.
So, what is the strategy for it to answer those issue is to use open source.
Andrew Froy will present you two tools that are based on open source.
So, Possible, which is one of the first projects that was started,
the Linux Foundation Energy Initiative,
and then what we can do, what we can build on top of Possible.
So, I'll let the floor to Joffre to present in detail the tools.
So, hello everyone.
So, the first project, Possible.
So, Possible, it means Power System Blocks.
Blocks, so this is a software component that we have as a foundation of many other applications,
especially at RT, so we have something like 15 projects that are based on,
as a list, a few components that has been developed in Possible.
So, what is the content of Possible? What is it?
So, this is many things, but it's the first way to model the Power Grid.
So, we have a data model that allows us to build, to have a green model
and to use it to make, for example, some evolution, some change for this grid model
and to study what will be the impact.
We also have some components for visualization of the grid
that will be integrated in some higher level application.
Also, what is very important is to be able to feed this data model with some data.
So, for that, we have some converter coming and to standard data format.
So, the most used one and the most famous one is the SIM, the SIM data model.
So, we have premium support for SIM converter, SIM data model support.
And also, what is very important is to have some interoperability with commercial tools.
For example, the one, there's two very widely used commercial tools
which has a PCC from Siemens and also PowerFactory.
So, we are able to import some data into our data model from these tools.
And also, we also have some converter from academic data format, for example,
MATPOWER, which is widely used for research and science.
And with this data model, we are able to run some analysis functions,
for example, powerful calculations, security analysis.
So, security analysis, for example, is a nice function that allows to test
what will be the impact of some contingency.
For example, we have a line loss, an outage on the grid,
and we want to see what is really the impact of this outage on the flow,
on the voltage to see if we have some trouble.
We also have some scientific analysis, short-circuit calculation,
which is also very important, and also dynamic simulation.
So, time domain simulation.
This is why we are integrated with another project of the Linux Foundation Energy,
which is Dainaw.
So, this is mostly written in Java.
And this has been designed to be as light as possible.
There is no dependency with complex framework or anything that takes decision
to how you are going to use it in a higher level application.
So, GridShoot.
GridShoot is an example of a tool that is built on top of the component of possible,
and that allows people to make some grid study.
And very different studies.
It starts from a real-time study, for example, security analysis,
to a long-term development study.
For example, we can with this tool, study what will be the impact of a connection
of a new renewable generation power plant on the grid,
and to assess that everything is fine if we connect this generation in a specific place on the grid.
So, this is a tool that has been moving to production very recently.
So, at the end of last year, since a few weeks,
and we still have some very early users.
And what we plan to have is 400 users in the coming two years.
And so, this tool will replace an existing tool,
which is at 30, since 15 years.
So, we have a team of more than 20 developers, and it is a growing team.
From the technical part, the technical stack that we are using is for sure 100% open source.
So, this is a micro-service-based application, very scalable application,
based on Java, Spring Boot.
We have everything based on REST API and also asynchronous messaging with RabbitMQ.
On the storage part, most of the micro-services are based on a PostgreSQL and Elasticsearch.
As it's quite difficult to manage such a distributed application with a lot of micro-services,
we have everything is deployed using Kubernetes cluster.
And on the front-end part, so this is the web application using our RAC.js.
And also, we use a little bit of WebGL for high-performance representation of the grid.
So, an important issue that we had with this is that
we have some Java component, which is very convenient to integrate it into,
I would say, a classical enterprise application where often the backend is based on the Java ecosystem
with Spring, Quarkus, or some kind of framework.
So, this is fine, but what we had is also needed to use these components
for high-performance also for research and data science community.
And most of the people from the data science communities are on the Python ecosystem.
So, the question was for us how to use the same piece of code in these two ecosystems
and how to share the code in Python and Java.
So, what we have done is to use another fantastic open source tool, which is the Gural VM.
And Gural VM, this is done by Oracle.
And this is several things, but we are using a component with native image that allows
to compile Java code into native code.
And thanks to this, we are able to build a C library for everything that we have impossible.
And with this library, we can build a classical Python extension module based on the C library.
So, some useful links.
So, for sure, there is a GitHub repository for both project possible and Gridshoot.
Maybe I can focus on the Slack channel, the two Slack channels.
So, see where is the place where we answer questions and we discuss with the community.
And also, there is an online demo of the application Gridshoot.
So, if you want to test it, you can do it.
So, we have an instance of Gridshoot that is deployed in the cloud.
And you can connect to this one just using, for example, your GitHub account.
Also, there is a YouTube video if you want to show a live demo.
So, this is a screen of the application.
What you can see here, just to explain what is it.
So, on the left side, we have the data manager.
So, starting from a case, from an initial, a green model, we have a way to create some tree of variant, of modification
that allows us to test different changes in the network.
And for all these variants, we can run some calculations and analysis
and then compare what is the best and what is the best one for us.
Then you can see on the right side that we have some way to represent, to display the grid.
So, this is a full representation of the French high voltage grid.
We have some substation diagram like here.
We have some what you call the network diagram, which is part of the grid that is shown in the bus, in the nodal view.
And then we can run some calculations.
We also have some table to see the data in a tabular form.
We have some specific user interface to show the results, etc.
So, this is one for the presentation if you have any questions.
Or if you want a demo of this tool, we can do it after the presentation.
If you are interested to have a more detailed view of this tool.
I was just wondering what format your network data is in.
And whether you could, for example, take in the open-street map network data and try analysis on it.
It's not complete, but can we do that?
So, this is not OpenStreetMap, this is MapBox.
But we can change the type provider to use whatever you want.
So, here we have used a very light type representation of the grid just to have a better view of the grid.
But we can use the OpenStreetMap.
How do you make the link between the grid and the end-consumption items or human on the grid?
Do you go to the machine-to-machine communication system so that the people stop consuming?
Or do you make advanced polls in order to know the consumptions within one hour, within one month?
So, this tool is a bit of a snapshot of the grid.
So, this is done by some other tool which are before this one.
So, we have the SkyDal, for example, which are doing the acquisition of the measurement.
And that has a database of the grid model.
And from this, we have some snapshots and then it can go into this tool.
Okay? I don't know if I answered the question.
How do you handle the stress when, for example, the grid is about to fold?
Do you cover any cases with humans at the end of the grid?
I don't know. Anyway.
Okay. So, we will answer the question later.
Thank you.