Facilitating growth: The TotalEnergies biogas business on AWS

Friends and colleagues. Good afternoon. My name is Samhita Shah and I'm Head of Marketing for AWS Energy and Utilities. And I'm responsible for marketing strategy for all our customers in oil and gas, power and utilities and renewables across the world.

Today, I'm excited to be here to talk about how AWS is working with our customers to reinvent energy and specifically how TotalEnergies is scaling their biogas business on AWS.

So today we will go over a few things:

1. How rapidly the industry is shifting and how innovation is truly required to enable this evolution.

2. The technologies AWS is providing to the customers to be able to accelerate innovation and reinvention. And you will hear from the TotalEnergies team on how they are reinventing energy on AWS and have ambitions in significantly scaling their biogas business.

3. And last but not least, my colleague, Nadia Bolinger, Global Head of Partner Solutions Architect for Energy. We'll talk about how together with TotalEnergies, we are building the foundation for TotalEnergies to scale their business and what architecture they're deploying to be able to achieve this.

The energy industry is the foundation of today's modern society. And what we're seeing are basically four key trends that are really shifting the landscape:

1. The rapid energy diversification. There are more fuel sources emerging such as biofuels, hydrogen ammonia and the list goes on and these different types of fuel sources are what's powering today's society. But more importantly, according to EIA report that just came out last month, non fossil fuel share of primary energy demand is going to grow by 34% by 2050. That's a rapid increase based on the fuel mix that we have today.

2. Just a significant amount of increase in electrification that's taking place. Global electricity generation is expected to increase by 76% by 2050. That electrification is mainly driven by EVs fleet electrification and home usage. And we're only going to see this trend rise. This significantly changing landscape also requires a more sophisticated energy portfolio management process.

3. Diversification. There has been a strong emphasis on this for many, many years. However, the urgency is growing and accelerating. Global investment in clean energy will reach $1.7 trillion by end of this year. And those clean energy technologies are really crucial for diversification and electrification goals we just talked about.

4. And lastly, customer expectations and how they are interacting with energy systems is rapidly evolving as well. As an example, EV sales jumped from 9% to 14% in the last year, that's 10 times increase since 2017. And that's just going to continue. Similarly, customer interactions with the power grid is evolving. Organizations are demanding more sophistication in managing power with distributed storage and generation and visibility that they are seeking and demanding are rapidly evolving as well.

All these things are driving energy organizations to really rethink their participation in the energy ecosystem and how they integrate into other parts of the value chain, and more specifically how they are utilizing AWS technologies to navigate and accelerate the energy transition, whether it's advanced analytics, AI/ML, high performance computing or edge computing.

Ultimately, the reinvention is to achieve five business outcomes that these energy organizations are looking for, whether to meet 50% increase in energy demand by 2050 whether to lower operational cost and be efficient and pass on the savings to the consumers, reduce emission across the energy value chain, minimize risk to assets and personnel or to maximize safety and security of the infrastructure. These are all very important factors in which the reinvention is being driven and really what they're doing is accelerating innovation and development of new technologies.

It is clear that many of the technologies that we need to meet the net zero goal by 2050 are not yet developed or commercialized. There's significant amount of expertise, capital that's required to scale these technologies. And most importantly, it's absolutely clear that no single organization can solve these challenges alone and it's going to require a global community of partners and startups to accelerate the energy transition.

And that's why we're very excited to continue the AWS Clean Energy Accelerator program. The program was launched to foster co-innovation through energy partnerships. And leading energy organizations like TotalEnergies have been collaborating with startups through this program to help solve their clean energy and decarbonization challenges.

As we know speed matters and clean energy technology entrepreneurs around the globe are emerging as some of the fastest movers inventing and reinventing shaking the status quo. TotalEnergies and 11 of their peers in the program are testing and embedding those technologies within their own organization to be able to more efficiently scale those technologies.

On December 7th, we will conclude the third cohort with an innovation showcase in Dubai around COP 28 where the startups partners policymakers organizations in within energy industry and beyond will come together to learn more and discuss about these new, these new technologies that are so important. If you will be in Dubai on December 17th, I really urge you to attend the event and be a part of this change.

Now, lastly, our mission is clear, it is to enable energy organizations to transform and reinvent the industry. And we're excited to say that we have leading energy companies like TotalEnergies working with us to achieve this.

Now let me introduce my esteemed guests from TotalEnergies, Delphine Gareau and Ricardo Pilla Cota and to hear about their transformation journey and how they were specifically reinventing energy.

Delphine Garre is the Chief Information Officer of the Gas Renewable and Power segment at TotalEnergies and is responsible for the information systems supporting fast growing business such as LNG renewables, power trading, and marketing and biogas.

And at the forefront of Total Energy's transformation into multi energy company, Ricardo Peta is the Digital Transformation Manager for Total Energy's biogas business unit and is responsible for the transformation roadmap and management of the development projects to support the businesses growth and performance ambitions.

Please join me in welcoming Delphine.

Delphine: Good afternoon, everyone. I'm really happy to be with you today speaking about TotalEnergies strategy. So more energy and less emission. This is the purpose of everybody working at TotalEnergies. We believe that the demand in energy is going to grow while we all have to tackle the climate change issue.

You may have noticed in 2021 Total, an old company in oil and gas decided to change its name, becoming TotalEnergies. It was a strong message sent to everyone that TotalEnergies is reinventing itself, reinventing energy and we are becoming a multi energy company.

You can see behind me the journey we are all embarked on since then. Oil is today's core energy and our ambition is to excel in low-cost oil energy and also reducing the CO2 emission of our facilities. You see natural gas. As the center of that slide, natural gas for us is also a key topic and we believe that is a transition energy that will grow in the future. But the future is at the right of that slide, it's electricity and we are bidding the core of a new system in electricity today in TotalEnergies.

So let me a bit more. Let me explain to you a bit more about that. We also believe that we want to become an integrated power company. And you may have noticed that we are a 100 year company and we have been successful, we believe we have been successful in oil and gas because of our integrated model and we want to reproduce that success in the power.

So you can see that we are building strong capacity in renewables. We have huge projects and huge ambition. We are going to complement that production with flexible generation. And also we definitely need to be the storage capacity to manage the intermittency of that energies. At the core of everything we still have the trading and the trading is key for us to manage the cap to capture volatility and also to maximize asset values.

And at the end of everything, of course, our objective is to deliver clean from power to B2B and B2C customers. We have huge ambition in the volume of the production we're targeting. But also as you may notice in the road shape of 12% by 2030. And this is a huge ambition for us, a huge message of transformation of our company.

So let's talk about it because I'm dealing with it the same way our company is transforming itself. We definitely have to transform our IT is internally in the company. Our IT strategy is quite clear. We definitely need to optimize and even to reduce our legacy system while we need to transform and to build the system of the new businesses. And it was obvious for us that the cloud was key in such a transition.

So that's why we built a strategic collaboration with AWS in 2021 and we had three main targets. The first one was flexibility. We definitely need flexibility during a transition and it is key for us. The second one, we were looking at more security in our IT systems and the third one and not the least innovation, innovation for IT and also innovation for our businesses.

So let's have a look, look to this at the starting point. In fact, you may notice that our on premise footprint was quite huge in 2020 because 80% of our application, we are locating on our own on premise infrastructure. But our ambition was strong because definitely we need to go fast and to be able to target huge figures in the 2026. Our ambition is to have 60% of our application in the cloud by 2026. And we are also convinced that we still need application on premises.

And this has led to a huge transformation within the company to do so. We were also convinced that it was not possible to go on that journey without having a strong program internally so that we can gather all the energies of the IT people, the people that are within our businesses, the people that are on in our affiliates. And this is the way that we design the program internally to definitely move the cloud in a very strong manner.

So I won't go through all the details. If you have any questions, I'll be happy to answer afterwards.

But we are today happy to say that we are online with our ambition and we are quite confident in the fact that we will deliver our target by 2026.

So now let's have a look to what is at the heart of our today subject, the biogas. You may remember at the beginning of my presentation, I told you that the gas what at the center of our energy transition. And moreover, we're trying to develop new kind of gasses, the green gasses and the biogas.

We have acquired facilities mainly in Europe during the last three years in France and in Poland. And we're happy to be able to have such asset in the company right now. We are aiming at going faster stronger, trying to be developed in other countries. And we are leveraging some key partnership in US in India and in Brazil to do so.

And Ricardo who is working in the biou gas bu we'll give you a bit of a flavor of a typical use case that we have and where we are happy to have collaboration with AWS. That is not a huge it legacy system, it's quite small. And here AWS enable us to easily, quickly bring new on our sites and things that we are able to scale.

So Ricardo, if you want to join me on stage, explaining the challenges that you had and how you face them with AWS.

Hello. So I'm Ricardo. Happy to be here with you today. I'll start by a quick explanation on what is the biogas activity?

Um if we start on top left, um we have the feed stocks. So it's basically the input for our, for our uh installations. The end goal is to produce methane and this methane is coming from basically organic waste. This organic waste is taken from um either at the agriculture uh activity or food industry or uh eventually, in some cases, the municipality's uh organic wastes.

And then the second step is more an industrial uh step. Uh they, they, they have sev several different steps. And the main actors of this second step are the bacteria. So bacteria are basically doing anaerobic digestion, transforming um the organic waste into uh methane nc un co2.

So we have the method that we want to value and sell to the market. And so the, the, the, the the commercial parts of the feedstock part, we need to make sure that we saturate those units. Uh we need to make sure uh so the volume of, of, of the, the the the the the organic waste that we are buying, we need to make sure that we are not paying too much. So then the activity is still profitable. And the last one is uh what are the type of organic matter that we are buying? And this is a commercial activity that is impacting a second constraint.

That is below, that is the biological uh balance. We need to make sure that those bacteria remain alive but also happy and healthy. So they can still continue producing methane and their functioning uh is a pure biological uh functioning and they impact eventually the the mechanical and industrial process.

So this is this is what we call the multi constraint uh for a relatively small to mid size uh industrial installation. It's a quite complex uh system to manage so complex systems to manage with multiple constraints. Having a data driven strategy is uh is is is quite fit, fit to purpose.

So the project that we are presenting today is focused on the 2nd and 3rd constraints of the bi biological balance and the operational limitations. Uh so to control and to and to manage them properly, we need to have the data of what is happening in the operations. Having the data, meaning means connecting the hundreds of sensors that we have on site and using this data in the right way. So we can properly manage the the operations.

So the context in the context of the, the division that i work is driving innovation and digital transformation in the bio gas uh uh business unit. There is a very small uh division into the in total energy uh group. So total energy is a very large co corporation with some uh very good point. It's a very resourceful company and uh sometimes we have a large amount of processes uh the short term value uh expectations and the scalability questions that are there that can be put.

And we, when we put all this together, launching an innovation project can be a challenge. So the key uh uh so the key for, for, for launching this project was finding a very small parameter that would still be meaningful. So we could uh iterate it afterwards. In our case, uh the very small perimeter was connecting one bargas site uh that was in south of france, it's called deion.

This uh this uh operational site was um struggling with connectivity problems and also underperformance for a couple of months. So uh having the data connected and, and exposed was a win win for them. And for us, we, we could test this, uh we could test this uh this uh this, this uh this project. So that was what we call the phase one.

And from the phase one, we have uh ff phase two and a phase three that is ongoing. If we take, take into uh take a look on the phase one, we basically took the data from sensors on site that were uh that were now on a local scada and pull this data into the cloud on a temporary uh um timestream database and exposed on a, on a, on a very simple dashboarding.

Um and this was phase one and phase two came to as a temporary and a testing phase uh on two main uh tests. A first one was basically um uh putting in place a second method for data acquisition uh and also putting in place a data normalization step. So this data no normalization step was very important to make sure that the solution would be uh scalable afterwards.

A third phase that is still ongoing is taking everything that we could gather from the first phases and the failures as well. Uh and building a solution that is defining a pattern for the next uh site integrations. So we have uh from the end of the first of the third pha phase, we will have a pattern uh regarding the data um architecture, the infrastructure architecture and the governance uh around uh the, the, the, the monitoring, the monitoring um uh theme if we, so this was uh zoom in, in the three phases.

If we zoom out from a project management point of view, we have um in phase three the eight sites uh in france that will be connected. Those data are pulled from the locals scada systems in a raw format. And they are normalized in uh in uh in sitewise component. And the raw and normalized uh data are stored in the first layer of data. It is called the sources data.

Um sitewise uh will also allow us to calculate some, some what we call a da data product. So it's data calculations. For example, if we want to know the, the cumulated production for the day, uh we can calculate in sitewise, this is not a pure source data, it's already interpreted. So uh we introduce, we will introduce a second layer of data that is called domains.

And the end goal of all this and the philosophy behind is to link those data and the use of those data to direct business uh goals.

Um in the end of phase three, we have three main, um we have three main use cases that will be that will be uh delivered. Our first one is uh the usage by internal applications. In the moment. we have one that is uh scheduling the production and planning the the operations.

A second use case is more operational and and more more focused on uh near real time monitoring. Um this is to avoid operational problems uh before they happen, uh trying to anticipate with more informed uh dashboarding uh typically to be used on site and also in a centralized control room.

And a third one is uh more performance driven. So either by uh a performance um follow up uh and reporting or uh more analytics. So, having um uh advanced dashboarding and um and mostly access to historical data.

So this this pattern here uh is delivered in the end of phase three. And what, what is holding uh for the future for this project is basically following the same pattern. So it means adding a new data source, uh structuring it in the data in the, in the cloud. So it can be used by uh and uh by, by, by the, by the uh direct use cases, business use cases.

So the vision for this uh project and the for the following phases is to have a multi source data like um fully connected to our um digital ecosystem.

Um and this means uh uh progressing step by step because it takes time uh first because from a development point of view, we have literally everything uh to build and second from an organization and the first from a change of management point of view, we need to make sure that the uh our progress is going in the same direction as the organization that is evolving.

So, on this journey, we're counting on the collaboration with the aws uh to, to, to, to keep supporting us on the on the on the next steps and to complete the the presentation and i invite nadia bollinger is um head global partner uh for infrastructure solutions for energy in aws

I think so, turning big ideas into innovation like the data biogas project at TotalEnergies requires experimentation using design processes like the double diamond design process shown here. Coupling design processes with the tools available in AWS, organizations like TotalEnergies can capture data, learnings and results from experimentation and then rapidly apply those learnings and iterate as needed.

Experimentation also gives you runway to ensure that you're building the right solutions for the right problems. And it allows you to build proof so that you can galvanize support from senior stakeholders and those who will ultimately be affected by the change.

The double diamond design process was developed by the British Design Council in 2004. And it's really rooted in experimentation. The two diamonds represent a process of exploring a problem statement widely and deeply which are called divergent thinking and then taking focused action which are called convergent thinking. The process highlights that innovation is not linear, that no idea is ever really finished. And that to get to the right solution is in fact a reflective practice.

So let's take a look at the double diamond design process and how we applied it to traverse the innovation journey with TotalEnergies.

In Phase One, as Ricardo mentioned, we explored the problem statement through proof of concept. We wanted to see if we could gather data from a site, store it centrally in some type of storage inside of the cloud and then do some reporting against it.

The proof of concept was run by our AWS Professional Services team and they used some of our AWS native services to run that proof of concept. They used IoT Greengrass to collect data locally at the site. They used DynamoDB to explore storing that data in the cloud. And then they used our own QuickSight reporting tool to do some reporting.

The proof of concept was successful and TotalEnergies was pleased with how quickly we could get it stood up. But we wanted to explore the problem statement at scale - what happens if you add multiple sites, does the problem statement change and if so how to manage multiple sites?

We added another service called AWS IoT SiteWise that allows you to manage industrial assets at scale. Plugging in a second site was successful and it gave us the confidence that we needed to feel like we fully understood the problem statement. And now we were ready to figure out what is the best way to solve it.

In Phase Two we explored alternative solutions to ensure that we had the right solution. There were a couple of areas that we wanted to explore some alternative solutions for:

1. Data acquisition
2. Data management
3. Reporting

For data acquisition, TotalEnergies wanted to explore some vendor solutions. For data management, they knew that IoT SiteWise was the right service to use, but they wanted to explore and experiment with different data models. And then finally, for reporting, they realized that they had different reporting needs - one for real time reporting and one for longer term use cases like analytics.

So let's take a look at where we landed after we explored these alternative solutions and ultimately found the best one.

What we learned from our exploration of the problem statement and through these alternative solutions was that there were really four key things that we needed to solve:

1. Although we were starting with 12 sites, we knew that eventually this solution would scale globally. So whatever we architected, we had to architect it so it could scale easily.

2. Each site was created differently - they came in with their own potentially OT system, own set of data models. So we needed to create a virtual representation of a biogas facility that each of these sites could be mapped into so that we could aggregate data holistically across the sites and then more easily integrate new sites as they came online.

3. As Ricardo shared, the data is quite complex. Not only is the biogas process itself very complex to model, but because this was a global operation, we had a situation where some sites were coming in with different units of measure, different labels. So we had to normalize the data into a standard format so that we could manage the data holistically across the different sites.

4. And last but certainly not least - security. Although this is a read only system from the process control environment to the business network, we have to ensure that we transfer the data in a secure way as possible.

Now, I've already touched on this but at a very, very high level, our architecture followed the standard extract, transform, load data ingestion architecture pattern with a visualization engine added at the end.

In this model, our solution is architected to align with these different parts of the process:

1. First we ingest or extract data locally from a site.

2. Second we transform that data into a high quality, reliable format that aligns with our reporting requirements and intended use cases.

3. Third we load the data into some long term storage like a data warehouse or a data lake.

4. And lastly we visualize the data and enable advanced analytics through reporting.

So here is the architecture of the final solution that we landed on:

Central to the solution for TotalEnergies is a family of services that we call AWS for IoT. AWS for IoT is built on the backbone of AWS cloud services, which includes a wide portfolio of over 200 services that can do data analytics, networking, compute, storage, artificial intelligence, machine learning - you name it, there's probably a service that can do it.

AWS for IoT also includes a set of services that allows you to collect, process, store, and analyze data from IoT devices. And what I really, really love about this architecture is that it exemplifies how leveraging AWS for IoT can really simplify how you build, manage and deploy IoT based applications.

So let's dig in:

In the previous slide, I mentioned that our architecture modeled the standard extract, transform, load process. So if we start from left to right, we begin by ingesting the data locally using an AWS IoT service called Greengrass.

AWS IoT Greengrass is an open source edge runtime service that allows you to collect data locally to site and then push that data into the cloud. It consists of two parts - a client edge software that gets installed on a local device or a gateway and a cloud service.

AWS IoT Greengrass enables you to collect and analyze data closer to where it is collected and generated. And the cloud service allows you to do data management and manage IoT applications across your entire fleet.

The IoT Greengrass service is also pre-integrated to a number of other AWS services including IoT Core, which is the next part in our architecture.

And AWS IoT Core is a message broker. So we already know from our exploration of the problem statement that there's two reporting requirements here - one is for real time or near real time reporting and the other one is for more longer term kind of data analytics use cases.

So we needed some type of a service that could come in and send data to both of those locations - one to the reporting engine and then one to long term storage. And that is exactly what IoT Core is designed to do.

Now at the center of this architecture, and something Ricardo already mentioned, is IoT SiteWise. IoT SiteWise is effectively what allows us to scale the solution globally.

IoT SiteWise is a managed service that allows you to model industrial assets and then ingest, process, store, and visualize data coming in from those assets. If you have a biogas or another type of industrial facility, you need to have a virtual representation of that physical asset and all of the characteristics that define that asset including what time series data is being collected and how that data is formatted.

If you have sites that are coming in with different units of measure or they have different labels, you also need a service that allows you to normalize that data into a common format so that you can manage that data holistically across your entire fleet. And this is where IoT SiteWise comes in.

IoT SiteWise serves as the foundation and the platform to model an industrial asset and then do something with the data coming in from that asset to produce insights.

Another key function that IoT SiteWise performs in this architecture is that it acts as a temporary database for the near real time reporting. As Ricardo mentioned, this is a key requirement or objective of this project and IoT SiteWise acts as a temporary database so that you can view the data in near real time and allow the sites, the central site, to monitor these assets with an acceptable and reasonable response time.

Now, after processing the data, we send it, I think on an hourly basis, to long term storage which we call Amazon Simple Storage Service or S3.

Now we come to the last part of the architecture - visualization. We've already talked about how TotalEnergies experimented with a number of different solutions here. Ultimately, they landed on using Grafana for doing near real time reporting. And that was mostly motivated by the fact that it can very easily integrate with the IoT SiteWise service.

And then they use Power BI for longer term analytics use cases. And then the last thing to mention is given that the data is stored in S3, it also opens up the opportunity for the organization to build internal applications as they see fit.

So as Ricardo and Delphine shared, you know, our journey is really just beginning and we're so excited to continue collaborating with them as we scale up the biogas facilities and they accelerate their journey to the cloud.

But as we reflect on where we've been, we thought there were a few key takeaways we could share with the folks in the audience, for those of you who might be trying to do the same thing.

So with any big undertaking, our first piece of advice is to start small. Experimentation is key to ensure that you do the right things. And then once you figure out the things that you want to do, that you do them right.

It can seem daunting to try and digitize a global operation with tens, hundreds of sites, thousands of sensors sending high frequency data. But if you start small, you can learn at a smaller scale and then you can scale that solution using services like IoT SiteWise.

The next thing that we want to share, and coming from the Partner Solution Architecture practice, is to lean on your solution architect. AWS has a full breadth of service options that allow builders to create solutions with the highest flexibility. But this isn't a journey that you have to go on alone.

Our solution architects are specially trained in how to optimally use our services to solve your most complex challenges. So please lean on them as you start this journey yourself.

Now, another key finding from our journey is that when you consolidate and centralize data, you're going to learn things and you're going to get insights that you can't even think about right now. So if you are considering taking that next step and centralizing your data, just know that when you do that, you are gonna start looking at your data in a different way and it could potentially transform the way that you run your business.

And last, but certainly not least, is security. So just know that you can in fact pull data from a process control environment in a secure way and view it in your business control environment or business network. And I can personally attest to how important security is - security is priority number one, having worked in the oil and gas industry for 20 years. And just know that you can securely pull data from the edge and put it in the cloud.

That wraps up our presentation and we really appreciate everyone coming in and seeing our presentation. We will take questions outside. So if you'll join me, let's give everyone a round of applause and end today's presentation.