Onde a subsuperfície instável se encontra com a infraestrutura de superfícies

Recent engineering accidents and natural disasters have highlighted across many industries the need for better interaction between subsurface and surface information to create more resilient and sustainable communities. In many ways we are in the early stages of a new way of thinking where Geosciences and Engineering are a continuum rather than separate disciplines.

The combination of both disciplines don’t come without challenges. Subsurface models are “pseudo” digital twins of nature, carrying different levels of uncertainty. Subsurface data is generally scattered, irregular, and furthermore, expensive to obtain. How can geoscientists effectively communicate the risk and uncertainty of subsurface models to engineers? What are the current technology trends (e.g., IoT) that can help measure, reduce and even manage the risk and uncertainty of subsurface models? Addressing these challenges will support better data driven decisions in geosciences for managing the life cycle of an infrastructure or asset.

Overview

[00:00:00.305]
(upbeat music)

[00:00:10.610]
<v ->Hello everyone.</v>

[00:00:11.800]
And welcome to Lyceum powered by Seequent.

[00:00:17.590]
My name is Ignacio Torresi

[00:00:19.070]
and I work as the Solutions Director

[00:00:21.280]
for geo statistics and Seequent.

[00:00:23.240]
And today I have the pleasure to be the facilitator

[00:00:26.790]
of this panel discussion, where we have some really,

[00:00:29.510]
really interesting professionals.

[00:00:31.090]
And the title of the panel discussion is

[00:00:34.110]
Where the Uncertain Sub-Surface

[00:00:36.780]
Meets the Surface Infrastructure.

[00:00:40.930]
So to start, I would like to introduce Ray Yost,

[00:00:45.330]
Principal Engineer of Advision.

[00:00:47.600]
We also have Jeremy Wrathall CEO of Cornish Lithium.

[00:00:52.630]
We have Christian de Santos CEO of SAALG

[00:00:55.800]
technology company from Spain.

[00:00:58.130]
And finally we have, Hiromasa Shima

[00:01:00.480]
Executive Officer of OYO Corporation.

[00:01:03.470]
I’m very, very thankful for them for their time.

[00:01:06.920]
For them to take their time to be here with me

[00:01:09.970]
in this really interesting discussion.

[00:01:12.135]
So, Ray first, can you please tell us

[00:01:15.370]
a little bit of your background.

[00:01:16.780]
<v ->I’m Ray Yost.</v>

[00:01:17.820]
I’m a geoscientist and geological engineer.

[00:01:21.220]
I spent about the past 30 years working

[00:01:22.850]
between the mining industry and consulting

[00:01:25.080]
to the mining industry in areas around design,

[00:01:29.560]
risk assessment, risk management.

[00:01:31.770]
Those are my principle areas of interest

[00:01:33.580]
as far as this panel, and what I bring to the panel

[00:01:36.210]
is this focus and interest in understanding uncertainty

[00:01:42.811]
and how that goes into and translates into

[00:01:45.270]
how we make decisions around design.

[00:01:47.640]
I’m looking forward to the discussion

[00:01:48.840]
and I’ll turn it over to my next panel member.

[00:01:50.950]
<v ->Hi, my name’s Jeremy Wrathall.</v>

[00:01:52.390]
I’m the founder and CEO of Cornish Lithium.

[00:01:55.730]
My background is I’m a mining engineer.

[00:01:58.210]
I trained in Cornwall in the UK,

[00:02:01.180]
and then I went to work in the industry for awhile

[00:02:03.550]
and then became an investment banker

[00:02:05.870]
financing mines all over the world.

[00:02:08.160]
And that was until 2016, when I started Cornish Lithium

[00:02:11.860]
on the basis that I was going to need an awful lot of lithium

[00:02:14.740]
for the transition to electric vehicles.

[00:02:17.780]
And that’s when I realized that

[00:02:18.690]
there was a good opportunity to do that.

[00:02:21.060]
I given some historic records

[00:02:22.540]
of lithium in Cornwall, so really.

[00:02:26.160]
And this is where my involvement with Seequent

[00:02:28.917]
and particularly Leapfrog comes in

[00:02:31.240]
when we started to become really heavy users

[00:02:34.550]
of Leapfrog software to enable

[00:02:37.310]
historical and contemporary data

[00:02:39.090]
to be put into three dimensions.

[00:02:41.320]
So if I’m a huge fan of what Seequent does

[00:02:44.800]
and really have used that

[00:02:46.720]
to power my company on development.

[00:02:49.560]
So that’s me.

[00:02:51.220]
<v ->So hi to everyone.</v>

[00:02:52.300]
My name is Christian de Santos.

[00:02:53.890]
I’m the CEO and one of the co-founders of SAALG Mechanics.

[00:02:57.520]
Basically my background, I’m a geotechnical engineer

[00:03:00.410]
with a PhD in solar mechanics.

[00:03:02.510]
So previously to my adventure at SAALG,

[00:03:05.260]
I’ve been working as a researcher for more than eight years

[00:03:08.560]
in my former university here in Barcelona.

[00:03:11.380]
So basically I was working on numerical modeling

[00:03:14.260]
and also trying to calibrate those

[00:03:15.980]
numerical models with monitoring data.

[00:03:19.040]
At SAALG Mechanics, what we are trying to accomplish

[00:03:22.280]
is to reduce geotechnical uncertainty by combining

[00:03:25.940]
medical models with geotechnical monitoring data.

[00:03:28.900]
So basically my experience is working in projects,

[00:03:32.740]
trying to make more robust numerical models

[00:03:35.590]
and trying to make better predictions

[00:03:37.520]
and understanding how this is all

[00:03:39.670]
going to behave during conception.

[00:03:41.800]
So really happy to be here and try to improve

[00:03:45.950]
and work with all the panels.

[00:03:48.350]
<v ->I started my career as a research geophysicist</v>

[00:03:52.740]
with oil corporation in 1980s,

[00:03:56.889]
the Japanese reading geo engineering farm.

[00:04:00.640]
Then I got promoted to the R and D manager,

[00:04:03.770]
then moved to California USA to manage several geophysical

[00:04:09.070]
instrument manufacturing companies there.

[00:04:13.160]
After I moved back to Japan early 2000,

[00:04:17.120]
I served as the head of the oil international operations,

[00:04:21.430]
then I’m currently managing the R and D center of oil.

[00:04:26.950]
I was active with building smart international,

[00:04:30.130]
and a director of Japan chapter of PSI.

[00:04:34.860]
I’m very happy to be here.

[00:04:37.040]
<v ->Thank you.</v>

[00:04:37.873]
Thank you very much, everyone.

[00:04:39.849]
Number one, I really like the,

[00:04:42.855]
the different mix of geographies.

[00:04:44.930]
You guys are basing all in different parts of the world,

[00:04:47.630]
which makes it really, really interesting.

[00:04:50.700]
But what I really like is the mix

[00:04:52.870]
of engineers and geologists.

[00:04:55.790]
And I guess that a lot of the time when we talk about

[00:04:59.580]
uncertainty of subsurface models,

[00:05:03.620]
me being a geo statistician, for example,

[00:05:05.910]
I always say about geo statistics, but actually today,

[00:05:10.870]
we want to talk about that from more of a infrastructure

[00:05:14.900]
point of view and the role of subsurface models on that.

[00:05:19.060]
So to start, I’m just going to give us a little description

[00:05:24.380]
of the panel today and geosciences

[00:05:28.850]
are more than ever important for the design project delivery

[00:05:32.500]
and maintenance of infrastructure.

[00:05:35.320]
Recent engineering accidents and natural disasters

[00:05:38.320]
have highlighted across many industries the needs for better

[00:05:42.590]
interaction between sub-surface and surface information.

[00:05:47.030]
And that is of utmost important to create more resilient

[00:05:51.220]
and sustainable infrastructure, therefore communities.

[00:05:55.720]
In many ways, we’re in the early stage

[00:05:57.730]
of a new way of thinking where geosciences and engineering

[00:06:01.580]
are continued rather than separate disciplines.

[00:06:06.110]
But the combination of both doesn’t come without challenges.

[00:06:10.070]
And I guess you guys all know that.

[00:06:13.200]
Subsurface models are still the digital twins

[00:06:17.060]
of nature and they carry different levels of uncertainty.

[00:06:21.990]
Subsurface data is generally scattered irregular

[00:06:26.320]
and expensive to obtain, right?

[00:06:29.200]
So, you know, how can we effectively communicate

[00:06:35.400]
that uncertainty to different stakeholders,

[00:06:39.080]
technology trends.

[00:06:41.420]
We believe in Seequent that technology

[00:06:45.200]
has a big role in addressing some of those challenge

[00:06:48.970]
and it will support better data-driven decision

[00:06:51.350]
in geosciences for managing the life cycle

[00:06:54.210]
of an infrastructure or an asset.

[00:06:57.510]
So without further ado, I would like

[00:06:59.900]
to pose some really provocative and interesting questions.

[00:07:04.880]
And I guess that we have Jeremy here,

[00:07:07.420]
and Jeremy, he’s a mining engineer.

[00:07:09.800]
I worked in mining for a long time.

[00:07:11.540]
I worked by having mining engineers as my clients.

[00:07:15.980]
So I would like to know from him is,

[00:07:19.246]
you know, from your experience,

[00:07:20.730]
is there a better ways or how can we actually better

[00:07:24.820]
communicate the risk and the nature uncertainty

[00:07:29.975]
of subsurface models to engineers

[00:07:33.620]
or for engineering purposes?

[00:07:37.560]
<v ->Well, thanks, Ignacio.</v>

[00:07:38.393]
Yeah, that’s a really good question

[00:07:40.960]
and really as mining engineers,

[00:07:44.159]
we actually do think in 3D,

[00:07:46.770]
but throughout my career, I’ve realized that most people,

[00:07:50.850]
most the public, particularly their perception

[00:07:52.920]
of reality stops at the pavement

[00:07:55.290]
or the ground beneath their feet.

[00:07:57.020]
They do not think at all about the subsurface.

[00:08:00.780]
And I think a lot of engineers of particularly

[00:08:04.800]
people who are building things on the surface,

[00:08:06.520]
don’t think about the subsurface.

[00:08:08.600]
And this is really where I like the way

[00:08:11.590]
you said we’ve got a new way of thinking.

[00:08:14.570]
We’ve got a new way of portraying the subsurface

[00:08:18.730]
by using software, such as the Seequent,

[00:08:21.850]
particularly Leapfrog, in our case,

[00:08:23.900]
to show in three dimensions

[00:08:26.600]
and indeed in VR, if necessary

[00:08:31.550]
engineer’s how the subsurface appears

[00:08:35.200]
and what are the risks and uncertainties of the subsurface,

[00:08:40.180]
which makes their job easier.

[00:08:43.060]
So in our case, we were taking,

[00:08:45.770]
you know, what do we know about the subsurface?

[00:08:47.510]
We know, where do we get that data from?

[00:08:52.030]
And really in our case,

[00:08:54.440]
we’ve got a huge amount of historical data.

[00:08:57.330]
In Cornwell where we were working

[00:08:59.500]
because my Cornwell was a historic mining center

[00:09:02.480]
for in fact, many, 4,000 years.

[00:09:05.410]
But obviously we don’t have data going back that far,

[00:09:07.250]
but we do have data going back at least 200 years.

[00:09:10.810]
And we can bring that remarkably accurate data

[00:09:15.020]
into a three-dimensional model, such as Leapfrog

[00:09:18.770]
and then show in live form

[00:09:22.920]
how the subsurface looks like

[00:09:25.360]
an MRI scan of the subsurface.

[00:09:27.690]
We can also bring in contemporary data like satellite data

[00:09:31.500]
and geophysics into that model to really augment

[00:09:35.630]
what is a really detailed

[00:09:40.660]
observation of the subsurface.

[00:09:42.890]
So I think it’s a really exciting time.

[00:09:47.330]
We couldn’t have done this even five years ago.

[00:09:50.230]
I don’t think that software that we’re using now exists.

[00:09:53.120]
Certainly computing speeds that we use didn’t exist.

[00:09:57.090]
We had to build, our first computer was actually

[00:10:00.420]
custom built with parts didn’t even exist in the UK.

[00:10:03.310]
We had to bring them individually from the U.S.

[00:10:05.760]
to get the sort of speeds we needed

[00:10:07.470]
to process this data quickly.

[00:10:10.410]
And I think really this is an extremely exciting time.

[00:10:14.847]
And I go back to Ignacio’s description,

[00:10:18.270]
this is a new way of thinking.

[00:10:19.780]
It’s so important.

[00:10:21.250]
And using Seequent software has really made

[00:10:24.520]
a massive difference to our project.

[00:10:26.180]
So I don’t know if anyone else has got comments on that,

[00:10:30.007]
but that would be mine.

[00:10:33.773]
<v ->Yeah, thank you very much, Jeremy.</v>

[00:10:35.800]
Very interesting.

[00:10:36.830]
I think Ray has some comments to bring in.

[00:10:42.470]
<v ->Yeah, no.</v>

[00:10:43.390]
I agree with Jeremy that the use of this software

[00:10:46.160]
to allow a very fast and easy way to visualize

[00:10:48.860]
what’s happening in the subsurface is incredibly important

[00:10:51.640]
because being able to see something is the first step

[00:10:54.110]
in being able to understand it.

[00:10:56.550]
But around the pieces with risks,

[00:10:58.130]
there’s actually two different components.

[00:11:00.810]
One, we have to talk about uncertainties with our model

[00:11:04.220]
and two, we have to talk about risk

[00:11:06.720]
and those are different discussions

[00:11:08.500]
I wanted to focus on the risk piece.

[00:11:11.100]
Risk I’ve found is you have to talk about it

[00:11:15.760]
in terms of what the impact is to your audience.

[00:11:18.770]
There’s so many times whenever as geological engineers

[00:11:21.150]
or geoscientists, we talk about things

[00:11:24.040]
in terms of geological engineering and geoscientists.

[00:11:26.410]
And a lot of our audience doesn’t really

[00:11:29.300]
always understand that vernacular

[00:11:30.820]
where we’re sort of speaking in a very specialized language.

[00:11:33.730]
So we have to make sure that we’re communicating

[00:11:35.810]
in ways that are meaningful to our audience.

[00:11:38.350]
So if we’re in mining engineering or mining,

[00:11:41.090]
I found that whenever I talk to mining engineers

[00:11:43.770]
about what I’m doing in geological engineering,

[00:11:45.990]
it’s about how those things impact their schedule,

[00:11:49.830]
production, access, budgets,

[00:11:53.390]
income streams, whatever it is that I can’t talk about.

[00:11:57.360]
And in fact, this is a rule that I’ve developed

[00:11:59.220]
as the first rule of communicating geological engineering

[00:12:01.520]
is don’t talk about geological engineering.

[00:12:03.740]
You have to talk about things that are meaningful

[00:12:05.600]
to the people that are hearing your presentation.

[00:12:08.370]
And it’s not to say that you don’t, Jeremy,

[00:12:09.890]
just more that I found it so easy for us

[00:12:14.260]
in our discipline to talk about things that we’re doing,

[00:12:17.340]
because we understand those.

[00:12:19.080]
And so we have to make sure that we’re translating

[00:12:20.810]
those things into things that our audience

[00:12:23.320]
actually understands and can grasp.

[00:12:28.540]
<v ->Yeah, no, that’s fantastic.</v>

[00:12:31.820]
Sorry, Jeremy, you go.

[00:12:33.670]
<v ->I was going to say, right, yeah, that’s great.</v>

[00:12:35.604]
Absolutely, and hence why I used the sort of MRI scan

[00:12:41.640]
as a way of trying to reach out to people

[00:12:44.840]
who do understand, you know, that you can

[00:12:47.650]
actually see into something using modern software.

[00:12:50.410]
And I think that the risk element is,

[00:12:52.670]
is extreme from right through

[00:12:54.900]
from the geo geological modeling side.

[00:12:58.579]
You know, one of our good examples of that

[00:13:00.750]
is that we were trying to model a granite contact,

[00:13:03.570]
which is hugely important obviously.

[00:13:06.430]
And this, this granite contact was very, very unpredictable.

[00:13:10.470]
And I’m trying to translate that into

[00:13:13.040]
where to drill for our drilling engineers

[00:13:14.720]
was the critical issue.

[00:13:15.710]
So you’re absolutely right.

[00:13:17.610]
Risk, these models enable people to assess

[00:13:22.100]
all sorts of elements of risk,

[00:13:24.260]
including bot models and economic models,

[00:13:27.390]
whereas that, that simply didn’t exist a few years ago.

[00:13:33.550]
<v ->Oh, that’s great.</v>

[00:13:34.383]
That’s great.

[00:13:35.216]
Actually, you guys mentioned quite a lot

[00:13:37.510]
of different technologies, software computational capacity,

[00:13:44.310]
and that takes me to the next question.

[00:13:47.300]
You guys have been in the industry,

[00:13:50.080]
traveled all over the world,

[00:13:51.510]
and required different technology

[00:13:53.760]
to push your business forward.

[00:13:57.030]
You know, we have Christian de Santos here

[00:13:58.680]
who actually works for a technology company.

[00:14:02.750]
So what are the current technology trends, for example,

[00:14:06.110]
that can help measure reduce, or even,

[00:14:08.660]
you know, better communicate that?

[00:14:12.560]
Could we just quickly go through the table

[00:14:15.000]
started by you, Christian,

[00:14:16.820]
what do you think are the current technology trends

[00:14:18.910]
that could really change and really make

[00:14:22.760]
that way of thinking that we mentioned at the beginning,

[00:14:25.330]
Jeremy, really come together of geoscience

[00:14:28.290]
and engineering actually talking to each other a lot better?

[00:14:34.060]
<v ->Absolutely.</v>

[00:14:35.020]
I think that the trends for the past five years

[00:14:37.980]
have been IOT.

[00:14:39.630]
So basically how to measure and how to get

[00:14:43.050]
a huge amount of data.

[00:14:44.540]
I think that’s the furthest step.

[00:14:45.960]
You need to understand the real response of the ground.

[00:14:49.760]
And we are really good at doing that.

[00:14:51.530]
We are basically gathering automatically

[00:14:54.340]
many data from projects or that are ongoing

[00:14:58.150]
also have access to historical data.

[00:15:00.943]
I think that the key point and the future

[00:15:04.340]
is on the analysis of that data.

[00:15:08.560]
I think that we need to get this data

[00:15:10.710]
and start analyzing this data

[00:15:12.890]
and having a real impact into our projects.

[00:15:16.190]
So for us, is this the trend and the future

[00:15:19.810]
is we have been gathering data,

[00:15:21.930]
we’ll have a storage in data.

[00:15:23.540]
We have a huge amount of data,

[00:15:24.910]
but we really are not analyzing the data at that point.

[00:15:28.800]
We need to keep analyzing data.

[00:15:31.260]
So from my experience, what we think is going to be

[00:15:34.940]
the future is combining those sophisticated numerical tools

[00:15:40.050]
that we are using for designing, also with the huge amount

[00:15:44.430]
of geotechical monitoring data that we have at that moment.

[00:15:48.310]
And basically, and reaching the quality

[00:15:51.130]
and the robustness of our numerical models

[00:15:53.360]
to make better predictions, reviews,

[00:15:56.067]
and your technical uncertainty,

[00:15:57.420]
and basically to better understand

[00:15:59.780]
how this all is going to help

[00:16:01.950]
during the actual construction.

[00:16:04.290]
I think that we reached the point.

[00:16:06.010]
We are good at gathering data,

[00:16:08.080]
but we need to get to the next step,

[00:16:11.100]
which is analyzing data, big data technology,

[00:16:14.800]
machine learning algorithm,

[00:16:16.280]
to make sure that we can get trends from historical data.

[00:16:21.180]
And I think that’s the future.

[00:16:26.310]
<v ->Thank you very much, Christian.</v>

[00:16:28.350]
And actually I would like to, to call to this question,

[00:16:33.861]
Hiromasa, and why might you call Hiromasa

[00:16:37.190]
is because yes, we do get a lot of information,

[00:16:41.260]
but when it comes to infrastructure,

[00:16:42.970]
sometimes it can be expensive

[00:16:46.500]
or even too risky to drill or to sample in a place

[00:16:51.190]
where you already have a building or you have an excavation.

[00:16:54.310]
Sometimes drilling is not possible,

[00:16:57.590]
or what we call direct methods are not possible.

[00:16:59.970]
And that’s when indirect methods come in.

[00:17:03.040]
And I know Hiro that you have worked

[00:17:04.610]
with geophysics for a long time.

[00:17:07.730]
So what do you think is the role of, for example,

[00:17:10.370]
you know, we have used geophysics for large oil

[00:17:12.820]
and gas projects that are 2000, 3000 meters deep,

[00:17:17.890]
sometimes even below the ocean, we have done it in mining.

[00:17:22.250]
Why no more in infrastructure?

[00:17:24.980]
Is there, do you think that’s changing?

[00:17:28.520]
Do you think there is a different way

[00:17:30.030]
of using geophysical information for infrastructure?

[00:17:33.320]
Can you tell us a little bit about that?

[00:17:37.640]
<v ->Yeah.</v>

[00:17:39.200]
Good point.

[00:17:40.780]
I think that the geotechical industry, civil engineering industry is

[00:17:47.630]
so much used to, to the art and they tended to think

[00:17:53.090]
several bore hole drilling is enough and tend to interpret

[00:18:01.710]
between borehole and borehole with a straight line.

[00:18:02.897]
So if we have to develop other area, if it’s there,

[00:18:08.920]
we have difficulty to make a drill hole, they just skip it.

[00:18:14.210]
And because they are not familiar with geophysics,

[00:18:18.610]
they don’t think about geophysics, that’s a reality.

[00:18:22.890]
So we need to educate people.

[00:18:24.970]
We need to educate the customer

[00:18:28.692]
and tell them geophysics is a good tool

[00:18:32.890]
to interpret the between the borehole and the borehole

[00:18:38.682]
and add the missed information between borehole.

[00:18:42.260]
But you know, the upper area, there is some difficulty

[00:18:47.642]
that we don’t have in oil and gas.

[00:18:49.520]
That is urban noise, pavement.

[00:18:54.588]
The cities are covered by concrete and asphalt,

[00:18:58.010]
and always noisy.

[00:18:59.600]
So we need to, develop some of the new technology

[00:19:03.210]
to overcome them, but again, yeah,

[00:19:07.900]
we need the geophysics to develop a better geotechnical or geophysical model.

[00:19:15.060]
That is our huge challenge.

[00:19:20.160]
<v ->Thank you.</v>

[00:19:20.993]
Thank you, Hiro.

[00:19:21.850]
And I guess that fits perfectly well into the next question.

[00:19:25.170]
I guess that sometimes justifying the acquisition

[00:19:30.320]
of new information is, it’s difficult.

[00:19:34.050]
We know it’s expensive

[00:19:35.950]
and sometimes you might get some resistance

[00:19:38.160]
by companies or by regulators for that.

[00:19:40.390]
So I’d like to go to Ray, or on his experience

[00:19:44.710]
in the educational side of geoscience

[00:19:48.270]
but you think enough is done,

[00:19:50.900]
not just to the engineer, right?

[00:19:52.950]
Because, Jeremy, being a mining engineer,

[00:19:56.153]
he’s close to us geoscientists.

[00:19:58.752]
He heard about geologists and geoscience

[00:20:01.900]
in the early days of his career.

[00:20:03.870]
But what about communicating geoscience to other type

[00:20:07.640]
of stakeholders, for example, like community

[00:20:10.430]
like government, and what about regulators

[00:20:14.630]
rights of those institutions that are supposed to

[00:20:19.450]
regulate and approve for construction?

[00:20:24.990]
Do you think that’s enough is being done

[00:20:27.140]
or we still have a long way to go to communicate your science

[00:20:30.464]
to that kind of stakeholder?

[00:20:35.090]
<v ->I think we know the answer to this one.</v>

[00:20:36.893]
I mean, just no, not enough is done

[00:20:40.660]
and it’s not anybody’s shortcoming.

[00:20:42.340]
It’s just it’s really speaks

[00:20:43.810]
to the inherent nature of their problem.

[00:20:46.170]
And which is kind this foundational language of uncertainty.

[00:20:49.310]
We don’t always have a good way to talk about uncertainty.

[00:20:53.290]
We have our ways as practitioners that,

[00:20:57.020]
that we think about uncertainty and deal with it.

[00:20:59.680]
But then we also have these people who are

[00:21:01.270]
either having to the stakeholders, as you mentioned,

[00:21:03.890]
or the decision-makers about uncertainty.

[00:21:05.830]
And they’re not, either don’t have that exposure,

[00:21:08.250]
they’re not comfortable with it a lot of times.

[00:21:10.650]
It’s been my experience that quite a few folks are really,

[00:21:14.690]
they live in this very deterministic world.

[00:21:16.980]
They don’t care if the answer can be between seven and 53,

[00:21:20.790]
they want to know that it’s 12

[00:21:22.070]
and it’s 12, and that’s the answer.

[00:21:24.890]
So we really need to think about how we,

[00:21:27.460]
what’s the language of uncertainty

[00:21:30.170]
that we can speak to a common language

[00:21:32.540]
that people understand what we’re bringing to the table

[00:21:35.350]
in terms of where we’ve gotten to

[00:21:37.320]
and the decisions we’ve made,

[00:21:39.050]
but also what that means in terms of things

[00:21:41.120]
that could still happen,

[00:21:42.480]
that we haven’t yet accounted for in our models.

[00:21:45.800]
And likewise, our decision-makers and the people

[00:21:48.770]
who are our audience, the receptors of all this information,

[00:21:52.200]
they have to be a bit more informed about

[00:21:54.260]
the inherent nature of how this stuff works.

[00:21:57.220]
I mean, it’s science, it’s engineering, it’s geoscience,

[00:22:02.290]
and there’s this expectation of certainty.

[00:22:04.850]
A lot of there’s a lot of reasons for that.

[00:22:06.350]
We could have a whole discussion about

[00:22:08.100]
why there’s those expectations,

[00:22:10.250]
but they really have to start to understand

[00:22:12.070]
that these things are not so to speak set in stone,

[00:22:16.870]
and come to the table with that open mind

[00:22:21.427]
on either side with both being able to explain

[00:22:23.970]
these things in a different language

[00:22:25.870]
and our audience understanding that it’s not definitive.

[00:22:31.747]
It speaks to some of the challenges in modeling.

[00:22:35.170]
And I’ll go in on a bit of a tangent in that

[00:22:38.280]
with these models, we’re often putting this hard line

[00:22:41.320]
in the model, to say, here’s this boundary

[00:22:43.390]
between this information and that information.

[00:22:45.750]
Where really some of the things we need to start

[00:22:48.400]
showing is a certain sense of fuzziness.

[00:22:51.140]
That’s okay, we think we know that this starts here

[00:22:54.367]
and we know that this is over here, but in between,

[00:22:57.360]
unless we have some sort of really good sense

[00:22:59.280]
that there’s a contact, we need to have a fuzzy boundary

[00:23:02.530]
and being able to talk about fuzziness

[00:23:05.570]
in what we understand is going to be a challenge going forward.

[00:23:12.180]
<v ->Thanks, Jeremy, I think,</v>

[00:23:13.560]
I believe you have a comment or a question.

[00:23:16.760]
<v ->Yeah, I think it’s a really good point.</v>

[00:23:20.660]
This question for non geoscientists

[00:23:23.370]
like community government to trying to understand,

[00:23:27.160]
and we are now talking about people

[00:23:28.800]
who have no concept of what lies beneath their feet.

[00:23:33.090]
And I think, you know, previously in order to,

[00:23:35.830]
for mining engineers particularly

[00:23:37.520]
to get that concept across,

[00:23:39.850]
we’d have to build a physical three-dimensional model

[00:23:42.140]
with a shaft made out of a piece of wood.

[00:23:44.670]
And there was no fuzziness.

[00:23:46.250]
It was exactly, you know, it was a physical thing,

[00:23:50.120]
but it wasn’t necessarily accurate.

[00:23:52.100]
What we can now do, and I’ve done extremely effectively

[00:23:55.200]
with our community in Cornwall and government

[00:23:59.800]
have shown them three-dimensional subsurface models

[00:24:04.160]
using in our case, Leapfrog,

[00:24:06.610]
which has been massively powerful.

[00:24:08.630]
They can see, say for example, in our case,

[00:24:10.940]
the ordinance survey map as the surface,

[00:24:13.320]
or even a physical rendering of the surface,

[00:24:16.370]
and then we dive through and show them the subsurface

[00:24:19.560]
and we can show the uncertainties of that.

[00:24:21.940]
We can show extremely effectively to people who are

[00:24:26.560]
non-technical what the subsurface it looks like.

[00:24:30.550]
I mean, sometimes they are a little bit baffled still,

[00:24:32.670]
but it is a huge step forward.

[00:24:34.990]
And, you know, building, I think, you know,

[00:24:36.730]
as we get into VR and whether people can actually

[00:24:39.830]
put on a headset and walk around the subsurface

[00:24:42.760]
will be even more powerful.

[00:24:43.830]
So I think there’s huge strides

[00:24:45.600]
being made in software right now,

[00:24:46.820]
which enables a non geoscientists

[00:24:49.450]
to actually understand the subsurface.

[00:24:55.790]
<v ->Thank you, Jeremy.</v>

[00:24:56.760]
Thanks, That’s that’s fantastic.

[00:25:00.857]
And I agree. It is a challenge when you said

[00:25:04.000]
that we knew the answer to that.

[00:25:07.739]
Definitely, definitely leaves something

[00:25:09.970]
that still a lot of work to be done.

[00:25:14.260]
I think that I also would like to call,

[00:25:18.440]
and I think Christian, the last time

[00:25:19.830]
we had a conversation and Hiro, you as well,

[00:25:24.260]
you guys refer to the challenges that we also have

[00:25:28.180]
on the regulation side.

[00:25:30.960]
So when you are in a bidding process

[00:25:33.560]
to win a construction contract,

[00:25:35.780]
for example, and you’re doing the piece of work

[00:25:38.110]
that is reducing uncertainty,

[00:25:41.530]
so a better calculation of the amount of steel

[00:25:43.850]
or anchors or safety there is required for a tunnel

[00:25:47.800]
that comes with challenges as well.

[00:25:50.410]
And sometimes reducing the uncertainty goes against being

[00:25:53.890]
the most competitive bidder from a monetary point of view.

[00:26:00.170]
So Christian and Hiro, and of course, open to everyone.

[00:26:04.000]
Do you have any comments on that?

[00:26:05.610]
That’s quite, I think that is a quite a lot of challenge

[00:26:10.050]
to make infrastructure not just more resilient,

[00:26:12.700]
but more sustainable as well.

[00:26:14.300]
So anyway, what is your experience with that, Christian?

[00:26:19.210]
<v ->Yes, absolutely.</v>

[00:26:20.740]
I think that nowadays asset owners and government,

[00:26:25.040]
has a major role on that, basically because at the end,

[00:26:28.580]
the designers and the general contractors

[00:26:31.100]
are designing and building up what they are asking

[00:26:34.340]
for with their specifications.

[00:26:37.170]
And if there is just the money,

[00:26:39.090]
the financial benefit is it’s difficult

[00:26:42.960]
to make better engineering, better construction.

[00:26:46.040]
So they need to put on the table aspects of sustainability,

[00:26:50.130]
as you said before, and not just the financial benefit.

[00:26:53.900]
I can explain you like my experience

[00:26:57.140]
being working in a real project,

[00:26:58.930]
where we use our technology to review

[00:27:01.770]
geotechnical and scarcity, and showing

[00:27:04.290]
the general contractor like the grant was better

[00:27:08.050]
and that they had the opportunity to review

[00:27:10.690]
the number of anchors and the general contractor tell us,

[00:27:14.690]
but I don’t want to change my design or my construction.

[00:27:19.300]
I have a really good price on the anchors

[00:27:21.450]
and what they want is to put as much anchors as I can,

[00:27:24.410]
because if I don’t put those anchors,

[00:27:26.460]
the asset owner is not going to pay me for that.

[00:27:29.470]
And that’s basically the contract.

[00:27:31.890]
We need to change the framework of the contract

[00:27:35.490]
to make easier for them to improve their design,

[00:27:38.530]
to make leaner designs, leaner construction,

[00:27:41.760]
and that’s the responsibility, or at least

[00:27:44.790]
they had a huge responsibility

[00:27:47.410]
coming from asset owners and governments.

[00:27:49.420]
We need to change regulation, absolutely.

[00:27:56.370]
<v ->Fantastic.</v>

[00:27:57.203]
Now I think that sustainability is super, super important.

[00:28:04.090]
Hiro you said you also had some experience with that,

[00:28:08.150]
especially Japan, which is a place

[00:28:10.320]
that geotechnically speaking is so complex.

[00:28:13.870]
Do you have anything to add about that?

[00:28:16.910]
<v ->Yep.</v>

[00:28:18.850]
So we need a specialist that know the underground

[00:28:22.900]
is complicated, but many people, usually people

[00:28:26.880]
doesn’t understand the complexity of underground.

[00:28:29.970]
They, we don’t see the underground

[00:28:32.690]
and we don’t understand the underground.

[00:28:35.320]
So we tend to underestimate the risk

[00:28:39.240]
of uncertainty for underground.

[00:28:41.730]
So for infrastructure, construction market

[00:28:46.290]
is kind of an open market.

[00:28:48.800]
That means the cheapest be there is the winner,

[00:28:52.520]
but these days we are now having a serious accident incident

[00:28:59.770]
caused by the uncertainty of underground.

[00:29:02.870]
So now government regulate to understanding

[00:29:05.960]
the quality of the geotechnical model.

[00:29:10.049]
And so I think two years ago,

[00:29:13.670]
Japanese government announced that most of the government

[00:29:19.560]
infrastructure project have to be designed

[00:29:23.980]
and managed by BIM building information modeling,

[00:29:27.390]
3D modeling by 2023, so in two years.

[00:29:32.550]
In two years, for the government infrastructure project,

[00:29:36.720]
we have to develop the 3D model,

[00:29:39.940]
and to communicate the risk

[00:29:41.330]
and uncertainty using a 3D model.

[00:29:44.800]
So the situation is changing and the people understand

[00:29:48.070]
that we have to pay more to reduce the risk.

[00:29:52.370]
So that’s a good deterrent.

[00:29:56.370]
<v ->Fantastic, fantastic.</v>

[00:29:58.620]
Okay, actually, that takes me to

[00:30:01.970]
you mentioned BIM, and we know that BIM

[00:30:06.400]
are generally used for the infrastructure

[00:30:09.740]
and they are called generally digital twins.

[00:30:13.910]
We had some 30 minutes or really interesting discussion,

[00:30:16.300]
but I want to throw out there the last question,

[00:30:18.790]
which is since Bentley has acquired Seequent,

[00:30:24.550]
there was a lot of conversation about, you know,

[00:30:26.700]
how Bentley managed and does it really well,

[00:30:31.570]
the digital twin world of infrastructure

[00:30:35.020]
and that Seequent,that capacity in the subsurface,

[00:30:38.810]
but, you know, I don’t want to get too deep

[00:30:41.140]
introduce statistics and talk about varigrams

[00:30:44.146]
and simulations, but I know you guys

[00:30:46.810]
will have different views on that,

[00:30:48.280]
but can we actually call a subsurface model, a digital twin?

[00:30:52.720]
Wouldn’t that be a risk in itself

[00:30:55.310]
to give a subsurface digital twin to an engineer?

[00:30:58.550]
He would probably believe that it is a digital twin.

[00:31:01.960]
So I generally like to call them partial digital twins

[00:31:05.760]
when it comes to a subsurface.

[00:31:07.640]
But anyone has any comments on using that term specifically?

[00:31:11.410]
Or would you just not use it at all

[00:31:14.490]
for subsurface, of course?

[00:31:18.850]
<v ->Well, if I can add something,</v>

[00:31:22.050]
I think that the reality is right now,

[00:31:23.730]
all of us, we are using digital twins

[00:31:25.750]
because it’s what people understand.

[00:31:28.530]
My perspective and my reality, and it’s funny,

[00:31:31.830]
it’s basically the people that don’t know how to use

[00:31:36.120]
or create a digital twin, they really believe in them.

[00:31:40.060]
They believe the fancy colors, the 3D models and so on,

[00:31:43.810]
and the ones that are behind those models,

[00:31:46.040]
actually building those models,

[00:31:47.550]
they know how difficult is to get consistent results.

[00:31:51.020]
So they don’t really trust a hundred percent.

[00:31:53.730]
They know that is a really useful tool,

[00:31:56.620]
but if it’s not used properly is extremely dangerous.

[00:32:00.070]
But yeah, I think these are twins is a nice word to use

[00:32:03.940]
and everybody likes it, but dangerous word too.

[00:32:09.089]
(laughs)

[00:32:10.070]
<v ->Okay, yeah and I agree.</v>

[00:32:11.850]
Ray, you have any comments?

[00:32:15.690]
<v ->Yeah, I mean, I agree in a sense</v>

[00:32:18.480]
that it’s a good starting point, to say,

[00:32:21.110]
this is this digital twin, but in a practical sense,

[00:32:24.730]
it does imply this duplicate and it had been applies

[00:32:27.910]
this model that’s going to capture

[00:32:30.120]
all these different things that can be happening

[00:32:32.400]
when really we know that models are simplifications

[00:32:35.270]
based on some sort of presupposed

[00:32:37.130]
characteristic of interest.

[00:32:38.960]
And so before we even started, we’ve decided what things

[00:32:41.650]
we’re trying to capture in that model.

[00:32:43.438]
And they’re going to rule out a lot of things

[00:32:45.370]
which may turn out to be important somehow later on.

[00:32:48.840]
So we can call it a digital twin,

[00:32:50.940]
but we have to also talk about what it’s a twin of.

[00:32:55.160]
It’s only a twin of certain characteristics.

[00:32:58.380]
The other piece I think with this is

[00:33:00.710]
that’s important to consider is this sample to volume ratio.

[00:33:06.340]
If we’re looking at infrastructure,

[00:33:07.580]
maybe that sample to volume ratio is reasonable.

[00:33:10.340]
In mining, it’s ridiculous in certain cases.

[00:33:14.660]
If you look at say the typical suite

[00:33:16.560]
of material strength testing that we have,

[00:33:19.240]
and you look at the,

[00:33:20.150]
just the volume of samples that we’ve taken

[00:33:22.120]
relative to the ground that we’re trying to characterize,

[00:33:25.150]
it can be on the order of one to a billion,

[00:33:27.100]
to one to 10 billion and saying that we’ve twinned something

[00:33:31.090]
with one sample from 10 billion

[00:33:33.750]
is really pushing the boundaries

[00:33:35.480]
of what we can talk about realistically,

[00:33:38.540]
as far as the digital twin.

[00:33:42.000]
Just to put that into terms

[00:33:43.500]
that people can understand a little bit better,

[00:33:45.910]
if you had 10 cubic meters sort of the size of a bedroom

[00:33:49.340]
or so, and you were trying to characterize everything

[00:33:52.090]
in that bedroom with one grain of sand,

[00:33:54.390]
it’s tough to say that that grain of sand

[00:33:56.680]
is going to represent a digital twin

[00:33:58.610]
of everything that’s in that room.

[00:34:00.690]
So it’s just, it comes down again to

[00:34:03.330]
some of the things we’ve been talking about with,

[00:34:05.840]
yes, we can use this as the term digital twin

[00:34:10.210]
as a way to describe what we’re doing,

[00:34:12.370]
but we have to communicate that uncertainty

[00:34:14.520]
and limitations as well.

[00:34:16.540]
<v ->Yes, thank you, Ray.</v>

[00:34:18.410]
I love the analogy of a grain of sand in the room,

[00:34:24.030]
but also the metric of sample versus the volume

[00:34:29.040]
that we’re trying to characterize.

[00:34:31.740]
We know that is a good way of trying to communicate

[00:34:35.190]
to people how reliable that digital twin can be.

[00:34:41.500]
Jeremy, do you have any, any comments on that?

[00:34:43.550]
And of course, Hiro, I know you want to say a few things,

[00:34:46.130]
so Jeremy, please, you can start.

[00:34:49.970]
<v ->Yeah, look, I think, thanks Ray</v>

[00:34:52.960]
that was an excellent explanation.

[00:34:55.240]
I think the word digital twin

[00:34:56.840]
is correct it as near as we can get

[00:35:00.320]
to showing non geotechnical people

[00:35:05.060]
what the subsurface looks like,

[00:35:06.560]
but we always have to tell them that

[00:35:08.930]
the information we have is imperfect,

[00:35:11.070]
but I would say that some information on the subsurface

[00:35:14.310]
is better than none because otherwise

[00:35:17.669]
it’s a complete blank.

[00:35:20.330]
And if we tell them that, you know,

[00:35:22.200]
that there is limited data

[00:35:23.320]
and there’s a lot of uncertainty still left to go,

[00:35:26.700]
then at least they understand a bit more

[00:35:30.070]
about that subsurface than they were before.

[00:35:32.240]
And also each dynamic with a digital twin,

[00:35:35.960]
you couldn’t build in more and more data as you get it

[00:35:38.780]
and improve it and make it more accurate.

[00:35:43.890]
<v ->Hiro, you mentioned you had some comments too.</v>

[00:35:49.370]
<v ->Yeah, I think the digital twin is somehow</v>

[00:35:53.140]
a little bit misleading to many people.

[00:35:56.640]
Usually people tend to think digital twin

[00:36:00.190]
is in fact a copy of underground,

[00:36:02.690]
but it’s not the true.

[00:36:04.360]
Underground, it’s so complicated

[00:36:06.730]
and we cannot make exact copy of underground.

[00:36:11.870]
But we need a model, and to make the decision on

[00:36:16.069]
to reduce the risk.

[00:36:17.610]
And I’ve heard the underground,

[00:36:20.243]
meets the purpose,

[00:36:22.350]
then I have to say that is a digital twin.

[00:36:25.687]
That’s my idea.

[00:36:28.410]
<v ->No, I love that.</v>

[00:36:29.770]
So it doesn’t need to be a perfect digital twin,

[00:36:34.280]
it just needs to be a model that is fit for purpose

[00:36:36.970]
of what you’re trying to do, either mining, oil, gas

[00:36:40.270]
or geotechnical engineer, so completely agree with that.

[00:36:46.470]
So I think that with this

[00:36:47.330]
we finished with our panel discussion.

[00:36:51.600]
Thank you very much for your time, everyone

[00:36:53.640]
It has been very engaging, very insightful.

[00:36:56.140]
Thank you everyone for the comments.

[00:36:57.901]
(upbeat music)