Geotechnical digital twins and BIM

Building Information Modelling (BIM) is now a familiar concept, but ‘digital twins’ are emerging as a phenomenon in civil infrastructure.

Drawing on the extensive experience of Jonas Weil in geotechnical engineering and Nick Nisbet in BIM development, this panel discussion explores the concept of the geotechnical digital twin and contrasts it to conventional digital twins and BIM. What are the differences and how they are related? What is the role of open standards vs. geotechnical specialist tools to enable success on infrastructure projects?

Overview

Video Transcript

[00:00:05.770]
<v ->So, welcome everybody to this Lyceum 2020 panel discussion</v>

[00:00:09.940]
on Geotechnical Digital Twins and BIM.

[00:00:13.980]
You’ll be familiar with BIM, no doubt.

[00:00:17.200]
It’s been a feature of civil infrastructure projects

[00:00:21.860]
for a long time now.

[00:00:24.290]
Digital Twins are much more new on the scene

[00:00:27.624]
and certainly a phenomena currently in infrastructure.

[00:00:32.610]
But of course, if you layer a geotechnical layer

[00:00:35.550]
across any of these things, then it takes on extra interest,

[00:00:39.780]
and perhaps the rules of engagement are slightly different.

[00:00:44.610]
And fortunately, I have a couple of great panelists

[00:00:47.200]
with me here today.

[00:00:48.720]
We’ve got Nick Nisbet,

[00:00:49.840]
who’s the vice chair of BuildingSmart in the UK

[00:00:52.970]
and CEO of AEC3,

[00:00:55.830]
been working on BIM for over 40 years.

[00:00:59.810]
And Jonas Weil, senior engineering geologist

[00:01:02.890]
at IC Group in Austria,

[00:01:05.030]
and Jonah brings over 10 years of provisional experience

[00:01:08.910]
in infrastructure and mining

[00:01:11.270]
with a focus on structural geology

[00:01:12.890]
in geological 3D modeling.

[00:01:15.620]
So without further ado…

[00:01:20.960]
Join us here, good to see you guys,

[00:01:22.960]
and really appreciate you guys taking the time

[00:01:25.200]
to join us for this panel discussion.

[00:01:28.610]
<v ->Thanks for having us.</v>

[00:01:29.665]
<v ->Also with us is my colleague, Gareth Christopher,</v>

[00:01:34.430]
who’s the business development manager for our EMEA region.

[00:01:37.820]
So, that’s great. Really appreciate the time.

[00:01:41.230]
And Nick, perhaps, if we started with you,

[00:01:44.640]
I know you’ve been involved with BIM for a long time.

[00:01:47.950]
So, how does where we are now

[00:01:50.720]
compare to the challenges you kind of encountered

[00:01:54.420]
at the beginning of your journey?

[00:01:57.640]
<v ->Oh, well, it’s been a long journey.</v>

[00:02:02.300]
I think all the professions were wedded to drawings

[00:02:12.420]
as the primary means of communication,

[00:02:15.435]
and I think each profession has claimed

[00:02:19.480]
that they think in 3D and draw in 2D,

[00:02:23.040]
and that was very much part of the education

[00:02:25.540]
of the architectural education I went through.

[00:02:28.410]
But little by little,

[00:02:31.880]
each of the trades and professions

[00:02:33.900]
had learned that actually delivering information in 3D

[00:02:38.810]
is beneficial to themselves and to the rest of the team,

[00:02:43.870]
whether it’s an infrastructure team or a building team

[00:02:47.400]
and whether it’s a construction or in use.

[00:02:50.640]
So, it’s been a gradual progression

[00:02:54.810]
taking each discipline as it comes

[00:02:57.650]
and making sure that the expertise is preserved

[00:03:04.920]
as we look at new ways of working.

[00:03:08.500]
<v ->So, is that really the ultimate purpose of BIM then</v>

[00:03:11.871]
to bring together these different disciplines?

[00:03:18.950]
<v ->Well, I think the aim is to have an understanding</v>

[00:03:26.040]
of the build environment,

[00:03:27.360]
whether it’s buildings or structures or geotechnical work

[00:03:31.050]
and tunneling,

[00:03:32.560]
and inevitably that means sharing information

[00:03:36.520]
to create the complete picture

[00:03:38.080]
because no one knows it all

[00:03:42.383]
and all the information has some relevance to everyone else.

[00:03:48.150]
And I think that’s the real strength of BIM

[00:03:50.270]
is that you share information

[00:03:51.650]
in a way that can be used in lots of different angles.

[00:03:55.540]
You can get schedules out,

[00:03:57.220]
you can get 3D views out, you can get animations out of it.

[00:04:02.350]
You can get all sorts of analysis from it.

[00:04:06.960]
And so,

[00:04:08.170]
it becomes what is now called the single source of truth.

[00:04:14.060]
Certainly a more coherent understanding

[00:04:17.930]
of the build environment than was possible

[00:04:20.180]
when we relied on separate drawings and separate documents.

[00:04:25.186]
<v ->And has it been hard</v>

[00:04:26.590]
to get people to kind of move their mindset?

[00:04:32.280]
<v ->Well, yes, it has been hard.</v>

[00:04:36.440]
I think I started in the late 70s, early 80s

[00:04:40.900]
when the task was to persuade architects

[00:04:45.270]
that their information was better in 3D

[00:04:49.470]
and better in BIM,

[00:04:51.790]
joining together drawings and documents

[00:04:53.720]
into something that was naturally coordinated and so on.

[00:05:00.940]
And then the conversation moved to structural engineers,

[00:05:04.597]
MAP,

[00:05:06.720]
and more recently,

[00:05:08.790]
I think civil engineers have had to go through

[00:05:12.950]
the same process

[00:05:13.840]
and there are a lot of other disciplines

[00:05:15.340]
I could mention too.

[00:05:17.070]
And I think in each case, it’s needed an honest evaluation

[00:05:21.950]
of what it was that we were trying to do

[00:05:26.100]
with the old drawings and documents

[00:05:28.380]
and what we can now do with BIM.

[00:05:33.030]
<v ->Right.</v>

[00:05:33.863]
Well, thanks, Nick.

[00:05:35.890]
So Jonas, from your perspective,

[00:05:38.080]
as a practicing geotechnical engineer,

[00:05:40.810]
engineering geologist, how does that compare?

[00:05:45.270]
<v ->It’s actually pretty similar.</v>

[00:05:48.770]
From my experience, we started with this way later.

[00:05:53.021]
(Jonas laughs)

[00:05:53.854]
So like 10 years ago,

[00:05:55.370]
we started with extensive use of 3D modeling

[00:05:58.450]
coming from mining.

[00:06:00.600]
But we realized that our colleagues who do the tunnel design

[00:06:04.270]
and who do coordination on construction sites,

[00:06:07.880]
they really appreciated to have the data together in 3D.

[00:06:12.120]
And the first applications that we had in infrastructure

[00:06:14.870]
was actually such cases, what I said.

[00:06:16.860]
We have difficult situation, complex geometry.

[00:06:22.080]
You have these tools from your mining project

[00:06:24.560]
that you showed us.

[00:06:25.820]
Can we not use it for the tunnel?

[00:06:27.120]
And we did things like that

[00:06:29.430]
that we just compiled for whole date

[00:06:32.230]
with tunnel documentation, with the design,

[00:06:35.310]
with existing knowledge on monitoring,

[00:06:38.130]
and they really appreciated it.

[00:06:40.470]
And the push came actually from our engineering colleagues.

[00:06:45.693]
<v ->Okay.</v>

[00:06:46.529]
<v ->So, we have quite a separation of roles.</v>

[00:06:49.180]
So, I see myself as a geologist mainly,

[00:06:52.560]
and we provide or we translate the ground foundations

[00:06:55.340]
for the engineers

[00:06:56.640]
and we have all working environments, and the engineers…

[00:07:01.260]
So, we are happy when we see they using our models,

[00:07:04.110]
and they get a 3D model and start making slices,

[00:07:08.170]
creating their own sections,

[00:07:10.950]
because this is again, a similarity,

[00:07:14.360]
many applications in infrastructure are related to 2D

[00:07:19.770]
or will always be because of course, if you have a tunnel,

[00:07:23.207]
you have a linear structure.

[00:07:25.030]
If you (indistinct),

[00:07:27.640]
you bring it down somehow to 2D depending on the scale.

[00:07:33.460]
But also for geotechnical applications,

[00:07:37.030]
the first quick approach

[00:07:39.340]
is often 2D calculation of a section,

[00:07:41.610]
might be a ton of cross section

[00:07:43.010]
or a section through a cut slope,

[00:07:44.960]
but this is most effective.

[00:07:47.800]
And if you would have a big uncertainty in your model,

[00:07:51.105]
yeah, especially in that case,

[00:07:53.640]
it’s a very good tool before you do any 3D analyzes.

[00:07:57.890]
So, but for us,

[00:07:59.990]
it’s important to point out

[00:08:02.530]
these 2D sections must always be derived

[00:08:05.870]
from the complete picture.

[00:08:07.340]
So, you need the comprehensive analyzers.

[00:08:09.590]
You need to understand the conditions in 3D,

[00:08:12.232]
and then you define where to put your 2D section.

[00:08:16.010]
<v ->Yeah, that’s a great point.</v>

[00:08:17.570]
And I like this idea

[00:08:19.450]
of the blending of the 2D world and the 3D world together

[00:08:22.490]
in the model.

[00:08:23.323]
It seems to make sense to me.

[00:08:26.305]
We do have one of your models here, Jonas.

[00:08:28.490]
So, perhaps if I just share that,

[00:08:31.490]
you could talk to us through the model

[00:08:33.660]
and what it’s showing.

[00:08:35.290]
I’ll just move here to that and start to play it.

[00:08:42.100]
So please, tell us what you’re seeing here.

[00:08:45.930]
<v ->With pleasure.</v>

[00:08:47.168]
(Jonas chuckles)

[00:08:48.001]
We shot this example at several locations.

[00:08:50.929]
On one hand, because it’s a very impressive project.

[00:08:54.570]
I mean, it’s 200 meter high-arched dam.

[00:08:58.460]
And with all the related subsurface structure,

[00:09:00.960]
and so it’s definitely one of our,

[00:09:03.610]
the upper end of the scale

[00:09:05.180]
in terms of infrastructure projects.

[00:09:07.570]
And the other hand is that we had a lot of time

[00:09:10.080]
and there was a lot of investigation going on

[00:09:13.960]
and we had a lot of data to put together.

[00:09:17.280]
So, here we see the final cut slope

[00:09:21.840]
of the left bank of the stem science.

[00:09:24.460]
This is from a photogrammetric model,

[00:09:26.520]
which we brought into Leapfrog

[00:09:28.750]
and we used it to document the slope documentation.

[00:09:33.060]
So, here we see the geological mapping of the cut slopes

[00:09:36.610]
as it is done in mine sites as well.

[00:09:38.950]
And we documented all of these cut slopes

[00:09:42.230]
over more than three years,

[00:09:43.690]
and we’ve put all of this data together,

[00:09:46.250]
but this was already at a later stage of the project.

[00:09:48.790]
This was when excavation had started.

[00:09:50.690]
Actually, we started modeling in 2013

[00:09:53.820]
when we had the first borehole data

[00:09:55.480]
and the first seismic profiles,

[00:09:57.930]
which this thing maps from tender documents.

[00:10:00.190]
And so, we started putting together everything we had,

[00:10:04.820]
and this continued throughout the project.

[00:10:09.610]
We can come back to this 2D, 3D or 1D (indistinct).

[00:10:13.831]
Here you’ll see a borehole with borehole geophysics.

[00:10:16.440]
There was a lot of in situ testing

[00:10:18.210]
with very different methods.

[00:10:21.140]
Here we see again, maps tunnel documentations.

[00:10:24.073]
This is not slopes, but all the tunnels,

[00:10:26.800]
all the access tunnel (indistinct).

[00:10:29.880]
We see geophysics.

[00:10:31.100]
This is again a example in 2D because we do profiles.

[00:10:35.750]
This too was a seismic investigations.

[00:10:39.270]
And then we use all of this data

[00:10:41.070]
that we have in a model for interpretations.

[00:10:43.670]
So, we created different kinds of models

[00:10:47.000]
for different purposes.

[00:10:48.665]
We modeled rock mass types for the dam design.

[00:10:53.913]
We modeled rock mass types for the slope design.

[00:10:57.820]
We modeled discrete discontinuities for several applications

[00:11:02.330]
because the slope stability was mainly controlled

[00:11:05.020]
by patterns and discontinuity patterns.

[00:11:09.580]
But all of these interpretations

[00:11:11.350]
have been done in the same model

[00:11:12.940]
and with the same background.

[00:11:14.710]
So, all the factual data was available

[00:11:17.800]
and everything was used together

[00:11:20.010]
to derive the different models.

[00:11:22.550]
And I think that’s one of the key aspects.

[00:11:26.090]
And this is the reason why we call this kind of model

[00:11:29.480]
a geo-technical coordination model.

[00:11:33.160]
Because this is for us a source of truth,

[00:11:35.690]
if you want to call it like this.

[00:11:37.110]
It’s the comprehensive overview of all existing data.

[00:11:41.320]
And from this, we make our interpretation.

[00:11:43.370]
So here, again, example 2D cut slope,

[00:11:49.190]
slope stability analyzes.

[00:11:51.940]
Before, the 3D model was mainly used for dam foundation.

[00:11:56.040]
And this video is like two years old now,

[00:12:01.743]
when the slope documentation had just been finished.

[00:12:05.680]
And since then, the model is still evolving.

[00:12:07.980]
So most of the cut slopes and tunnels are done,

[00:12:12.530]
but construction is starting and monitoring is ongoing.

[00:12:17.210]
Routing operations are ongoing.

[00:12:20.231]
(indistinct) analyzes need to verify the model

[00:12:25.100]
because there might be some critical areas.

[00:12:28.240]
Additional tunnels are being built.

[00:12:29.600]
So this is really a living model

[00:12:31.530]
and it will stay with us

[00:12:34.040]
at least until the end of construction,

[00:12:36.470]
but then we will definitely hand it over for operation.

[00:12:41.680]
<v ->It’s amazing.</v>

[00:12:44.050]
You call it a geotechnical coordination model.

[00:12:48.188]
Do you think

[00:12:49.160]
we’ve really just been looking at a digital twin there?

[00:12:53.180]
<v ->Yeah. You could call it like this.</v>

[00:12:56.170]
It might be different

[00:12:57.449]
to what people from the building sector have in mind

[00:13:01.370]
when they talk about the digital twins

[00:13:03.170]
or from other engineering disciplines,

[00:13:10.000]
because they usually know exactly how their parts

[00:13:13.740]
and how their building look like,

[00:13:15.050]
and they have perfect 3D geometries of every screw.

[00:13:18.460]
We don’t have it.

[00:13:19.370]
We have assumption about the ground foundations

[00:13:21.470]
and we have documentations on the surface or in boreholes

[00:13:24.990]
or at geophysical profiles.

[00:13:27.108]
But what we do is we bring it all together.

[00:13:30.000]
So the twin is, yeah,

[00:13:32.480]
it still is schematic sketch of the twin,

[00:13:35.150]
but it includes at least all available information.

[00:13:38.200]
So, it’s the closest you can get to a digital twin

[00:13:42.944]
when you don’t know about the exact conditions of structure.

[00:13:47.640]
<v ->No, that makes sense.</v>

[00:13:48.870]
That makes sense. Thanks.

[00:13:50.230]
And Nick, is that consistent with your perspective

[00:13:53.270]
on what a digital twin is or should be?

[00:13:56.920]
<v ->Yeah, I mean, I think you’ve got the process</v>

[00:14:01.310]
of information gathering,

[00:14:03.570]
whether it’s a site survey

[00:14:05.160]
or the more sophisticated geo-technical techniques,

[00:14:11.140]
and so on.

[00:14:12.206]
Out of that, you get an interpretation,

[00:14:14.940]
which is the thing that can be really shared

[00:14:18.029]
along with the implications.

[00:14:23.660]
So, whether it’s a digital twin for a building

[00:14:26.920]
or for the rock structures around the dam or whatever,

[00:14:31.150]
I think the important thing is that

[00:14:34.120]
it always involves a degree of simplification.

[00:14:37.231]
And I think the other key thing about a digital twin

[00:14:42.281]
is the idea of a feedback loop

[00:14:48.260]
that what distinguishes BIM from a digital twin,

[00:14:52.900]
certainly in my view

[00:14:55.040]
is that it’s a two-way flow of information.

[00:14:58.910]
So you may well have monitoring of the geology

[00:15:05.020]
or the structure going and updating the model,

[00:15:11.460]
and the implications of that being fed back

[00:15:17.240]
to affect the way the building or the structure is operated

[00:15:20.940]
or investigated further.

[00:15:24.160]
And this was something that was emphasized

[00:15:26.550]
with the work we did with the Japanese

[00:15:28.780]
when we were working on the geo-technical models,

[00:15:33.940]
because they are very keen

[00:15:37.830]
that there is this fourth stage of monitoring

[00:15:42.260]
given the instability of a lot of their ground, and so on.

[00:15:46.330]
So, we ended up with a data collection,

[00:15:48.760]
interpretation, implications,

[00:15:51.420]
and then a fourth stage of monitoring.

[00:15:53.810]
And I think that’s that fourth stage

[00:15:55.331]
really makes a digital twin.

[00:16:00.260]
<v ->Yeah, that makes perfect sense.</v>

[00:16:01.840]
So how have those kind of critical differences

[00:16:05.360]
in the geotechnical digital twin,

[00:16:07.280]
the interpretation side of it

[00:16:09.714]
and this kind of, yes, the ground is maybe constant

[00:16:17.340]
but your understanding of it is always evolving

[00:16:20.701]
and that you have a lesser or greater degree of certainty

[00:16:25.960]
in your understanding of it.

[00:16:26.910]
How has that impacted the development of the IFC standards

[00:16:32.060]
for geotechnics?

[00:16:34.360]
<v ->Well, we decided to focus</v>

[00:16:38.460]
on the sharing of the interpretation,

[00:16:40.630]
and we wanted to make sure

[00:16:42.378]
that if that information joined up with other information

[00:16:48.210]
from the engineers,

[00:16:50.400]
that there was no misunderstanding.

[00:16:53.080]
And so we wanted to make sure there were attributes

[00:16:56.780]
that would describe the uncertainty

[00:16:58.580]
attached to any of the objects that were being shared.

[00:17:02.500]
<v ->Thanks, gentlemen, for your time.</v>

[00:17:03.910]
And we look forward to the questions

[00:17:06.730]
that people have put in the chat.