Lyceum 2021 | Together Towards Tomorrow
This presentation explores strategic approaches for overcoming common challenges associated with modelling large multi-disciplinary datasets for offshore windfarm development.
The challenges include managing and working with ‘big data’, collaborating across multiple disciplines, tracking updates and changes throughout the project cycle, and bringing it all together in a single integrated model.
Customer Solutions Specialist, Seequent
Customer Solution Specialist, Seequent
<v Nikhil>Hi, I’m Nikhil Sen.</v>
<v Becky>And I’m Becky Bodger.</v>
We’re both customer solutions specialists here at Seequent.
<v Nikhil>My focus is Central,</v>
our data management cloud solution.
<v Becky>And mine’s geophysics</v>
and the Geosoft line of products like Oasis montaj.
Thanks for watching the talk today,
I really hope you’ve enjoyed Lyceum so far.
<v Nikhil>This talk is a case study</v>
of a multi-disciplinary approach to managing big data
for offshore wind farm development.
<v Becky>The focus these days on green energy</v>
and renewable resources,
and the energy transition happening in oil and gas,
it’s not surprising that many of our existing customers
and new customers have approached us
to inquire about Seequent’s capabilities
for offshore wind farm development and ground modeling.
About a year ago, we decided to put a team together
and see what we could come up with in this space.
Our first model used open-source
drill hole data from oil and gas,
and I think it extended down to the Moho.
The customer said that while she understood
the concept we were trying to demonstrate,
they were really only interested
in the top hundred meters of the seabed,
so we knew we had to go back and do some more digging.
We ended up on the Dutch Government website,
looking at an entire project’s worth
of prime wind farm data.
And for the cost of an external hard drive
they would send us everything relating
to the Hollandse Kust Zuid wind farm zone.
So, first and foremost,
we’d like to acknowledge the Netherlands Enterprise Agency
for releasing the Hollandse Kust data package
under a creative commons license,
which has formed the basis
of this integrated modeling study.
The hard drive we received was three terabytes
of multidisciplinary data.
One of the early struggles
was how do we manage this much data?
How do we make it available when we were sitting
in different offices around the globe?
We had to do some prep.
We were sent raw and processed data,
and we needed to decide what to use for the modeling.
So how do we ensure everyone was using the same files?
Along with Leapfrog Works,
we decided very early on that we would need Central.
<v Nikhil>That makes a lot of sense,</v>
Central can do a lot to help with those problems.
For those of you that aren’t familiar with our solutions,
let me tell you a bit more about Central.
Central is a cloud hosted solution
where individuals and teams can share, audit,
and track their data as their projects evolve.
The secure environment is inclusive,
as it is designed for collaboration
across a variety of teams and specialists,
management partners, and stakeholders.
Central allows users to visualize
and interrogate 3D data outputs,
which are modeled and published from Leapfrog.
The visualization platform comes with a communication tool,
and is an ideal space to communicate change
and carry out the peer review process
after which decisions can be made.
Central’s audit trail also allows members
and managers of their respective projects
to not only track model updates from Leapfrog,
but also updates to source, and additional project data
stored and backed up in the data room.
<v Becky>From a very high level,</v>
what we’re looking at on the screen
are a few of the goals and decisions we wanted to make,
or get out of our integrated
offshore wind farm ground model.
So, you have the model at the center,
we’re feeding it a bunch of data,
we’re basically building our subsurface digital twin.
We’re doing analysis, further interpretations,
there’s a lot of collaboration happening
across different teams and stakeholders.
We’re making decisions,
and then we’re going out and we’re collecting more data
and we’re feeding it back into the model
and doing everything all over again,
we’re continuously updating this model.
And really this model is being used,
it’s going to inform our foundation design.
We’re making a lot of decisions
and a lot of review processes happening,
and of course, we’re narrowing our scope of investigation
and we’re planning future data acquisition
over smaller areas.
And the data only gets bigger and more detailed,
higher resolution, as we iterate and we create
a second version, a third version of the same model.
<v Nikhil>Central is adaptable to teams</v>
and individuals of any size and location.
Every project can be composed of members
with a variety of specialties.
Furthermore, you can tailor Central
to fit your project structure,
and this would cause a bit of forethought and planning.
<v Becky>And planning’s exactly what we did.</v>
My colleague Fiamma and I,
Fiamma’s a geotechnical engineer,
we sat down one day in the office,
and actually because of COVID,
it was the first time I was meeting Fiamma face-to-face.
So it was really nice to return to some normal practices.
We sat down though, and we sketched out
how we were planning to attack this ground model.
So we sketched it out,
we had a general idea of who needed to get involved.
We knew exactly the data we had,
and of course we knew what the end goal was.
We then took the sketch and we digitized it,
and then we added it to a shared ideas board.
And, you know, it looked something like this in the end.
And based on the data and who we had on the team,
we knew that there would be a GIS branch,
we’d have a seismic branch.
I think we had something like 500 SEG-Y, SEG-D files.
We had CPT, so that could all go into a geotechnical branch,
and then we knew that some of the initial work
we wanted to do was geostats,
and we had another person who could work on that,
so we would create a separate branch just for her.
And we have data sitting in data rooms
down at the very bottom,
and that’s going to feed into all the different branches.
We have another product called OpenGround
that would also store some of the source data.
And once we had this structure in place,
even though it looks very siloed,
with Central it gives us the option and the ability
to have visibility across the different branches.
So in the seismic, while Oli, our geophysicist,
is working on his seismic geological model,
he could actually bring in the CPT model
that Fiamma’s created,
or some of the data and the other geophysics,
or the other GIS stream,
and use it while he’s building his seismic.
So there’s that ability to visualize and see the results
from each of the different branches.
And just to put into a little bit of context,
we had Aaron sitting in Australia
working on dune migration,
or sea bed mobility analysis of the bathymetry.
We had Oli working on the seismic in the UK.
I was sitting here in the UK working on some of the mag
and the side scan sonar data and the other GIS.
We had Thom in the Netherlands
really helping us with our source data and organize.
And we had Fiamma working on all the geo-technical stuff,
and we had Carrie working on the geostats
and creating those probability models.
And so, you know, there was a big team
across different time zones,
really all contributing and working together.
And, in the end,
we ended up with something that looked like this.
So this is actually a screenshot from Central,
and it’s just showing the branches,
and you can just see how we planned it
and we were able to stick to it,
and we ended up with a pretty cool structure
and a pretty good integrated model.
But let’s actually jump into Central at this stage
and show you what it looks like live.
Let’s start with how we actually manage
the input and source data we received.
We were dealing with a lot of data
and really big individual file sizes.
I spent a lot of time going through the raw,
the processed, and different versions of everything
to find the best examples to use in the model.
And so for that, once I figured out what we were using,
I organized the project in the Data Rooms in Central.
<v Nikhil>The Data Room is a secure environment</v>
for organizing source data.
It also allows you to track the version history,
and is available to all project members
who can rely on a single source of truth.
<v Becky>So you can actually see,</v>
what we’re showing you here is actually the structure
and the files that are currently being used in our project.
So you can click on any one of these things,
like the bathymetry.
We can click on the raw file that we have here,
we can see the dates,
and if there’s a new version we can upload that
and it will get pushed right through to Leapfrog
and notify you that there’s a new version available.
The next issue we had, of course,
was team members working at the same time.
So with six colleagues working across three time zones,
and very busy schedules,
it meant that we needed flexibility
to work on the projects whenever we were available.
So how were we going to coordinate that?
<v Nikhil>The events page tracks all action items</v>
associated with the project,
and as you can see here,
we’ve got a number of different colleagues
from around the world who have been working on this project.
<v Becky>Yeah, at all times of the day,</v>
so it’s 2:00 a.m., 9:00 p.m., 9:00 a.m.. (laughs)
<v Nikhil>As you can also see,</v>
the project is divided into branches.
So if I go back to our history tab over here,
and we can have a look at this over here,
the project is divided into branches
based on the discipline,
before realigning with the master branch.
This allows multi-simultaneous access
and development of the project.
The other thing that branch structure really supported
was managing the overall size of each workspace.
We didn’t all need to work with 500 SEG-Y files,
Oli could keep this himself in the seismic branch,
and then choose what to align with the master branch.
<v Becky>There was also a lot of uncertainty</v>
in what we were trying to do.
We had a concept, but needed space to experiment
and test different methods,
try different things out, without overwriting,
or sort of contaminating each other’s work,
and so the branching structure really helped with that.
Each one of these nodes is a project
that’s been published from Leapfrog,
and so you can see there’s multiple nodes
in each of the branches.
These are different versions of each of the projects.
In order for it to be a truly integrated model,
we also need visibility into each other’s results.
So what we can do is, in Leapfrog,
you can right click on your geological model folder,
and click import geological model from Central.
And you’ll see that you don’t only have access
to the branch that you’re currently working in,
so that node that you downloaded, or that revision,
you actually have access to any of the geological models
in the entire project,
so that’s across any of the branches.
So for example here, if I click on the seismic analysis,
here are two of Oli’s seismic geological models,
and I can bring those into whichever revision
I currently have open in Leapfrog.
When you start sharing these geological models
all over the places, across multiple branches,
you really need a mechanism in place to manage the updates.
So if we go back, if Oli for example,
goes back and updates one of those seismic geological models
in his seismic branch,
how is that going to carry across to the model
that we’ve imported now into the master branch,
or into the geo-technical branch?
<v Nikhil>This data which has been brought in</v>
from another branch, or the data room,
over here with boreholes, creates a link,
and this means once a source destination is updated
with newer data, we see this little clock call symbol,
and that means new data is available.
And with a couple of clicks,
we can pull that in from Central,
which means we’re working from a single source of truth.
<v Becky>So that was the plan,</v>
that was how we managed to implement it,
and hopefully we were able to demonstrate
how effective Central has been
in building our integrated ground model.
And the final piece is using Web Vis in Central
to visualize the model itself,
and start a visual conversation and peer review process.
<v Nikhil>As you can see here in the project overview,</v>
in order to access Web Visualization,
we simply click on the revision number.
And furthermore, I can access the saved scene.
<v Becky>This is a really great way</v>
to share your modeling and your data,
and anything you’re doing,
with stakeholders who aren’t really familiar with Leapfrog.
<v Nikhil>And within this visualizing tool,</v>
we can actually start conversations
with our fellow colleagues,
managers, partners, and even stakeholders.
We can do this by creating a comment.
Just create a comment,
orientate our model,
pick a zone.
Furthermore, we can tag our fellow colleagues
who are part of this project.
This will then send a notification
to our colleagues involved,
and in particular Fiamma,
that a change has been highlighted
and that further action needs to be taken.
We can access our notifications
by clicking on the little bell icon here,
and you can see here, for example,
we have been mentioned in a specific comment,
and it’s available, and so forth.
<v Becky>Another thing you can do,</v>
just to show you how that works,
if you clear the scene and then reload it.
And if we click on the comment here,
it will update to the exact location
where we added our comment.
<v Nikhil>The important point to note here</v>
is that we are starting a visual conversation
which notifies our colleagues,
and that feedback loop continues
so we can start and complete
the peer review process before making decisions.
<v Becky>And we’re taking it out of email, right?</v>
So we’re making it visible to everybody,
and it’s really important to move that conversation
out of that closed email thread
and into the data itself where everyone
has visibility over it.
<v Nikhil>So that’s those users</v>
that are part of the project,
regardless of the role or location.
<v Becky>So that brings us to the end of our presentation,</v>
and we wanted to leave you with just a few key takeaways,
or lessons that we’ve learned along the way,
and the first is to make a plan.
The model we just looked at was not our first,
our first model felt a bit disorganized,
and we actually undertook a rebuild
of the entire project a few months ago.
That was a really big decision,
but we felt that we needed a better focus
on planning early on,
and that really significantly sped up
the entire modeling process for us.
<v Nikhil>The second take away is good housekeeping,</v>
garbage in, garbage out.
You need to curate to organize the data
from start to finish,
since with every stage of the wind farm development
the amount of data grows exponentially.
Finally, use the tools available
for clear and concise communication.
<v Becky>So thanks again for listening to Nikhil and I.</v>
Please feel free to reach out
to any member of the Seequent team
if you have questions about Central
or anything else you’ve heard today during the Lyceum event.
<v Nikhil>Have a great day.</v>