IIoT in heavy industry: mining, oil, gas, iron and steel – L&H Industrial sets new standards with 5G monitoring, tracking and IoT core

You talked about the gateway – maybe we can shimmy a little bit from the bottom up into the cloud. Start with the hardware and then move into the evaluation via data acquisition and data processing. What hardware did you need?

Ben

Of course, we looked around in the market again to make sure that we weren’t building just any industrial gateway, just for this application. But it became clear pretty quickly, it also has to be able to connect to 5G. We are now very quickly approaching the 5G rollout. This is a super exciting topic because we then combine edge with cloud. This also additionally has a global standard. We also have it downward-compatible to 4G, because the rollout will only really start to cover the whole country in the next few years.

IO-Link has been a must criterion for this to be plug-and-play – then you end up with IO-Link very quickly. We compare this a bit with USB for the mouse. IO-Link is the counterpart in the sensor area. So you screw on a vibration sensor, a temperature sensor, and virtually seconds later you see the measured values in the right scale, in the right format, and so on. This all runs in the background with the IO-Link protocol.

And that is then passed back to the network via a specific protocol so that it is available in the AWS IoT Core?

Ben

Right. We have selected AWS. We use the MQTT Hub from the AWS IoT Core. We send the sensor data via this interface. The photographs, pictures, run on a different protocol.

That means we now have relevant sensor data for the customer, we have cameras … we have data about the bucket itself, for example, or also operating hours. They are passed on with a service, so to speak, and via your gateway I have the option of routing this data out via appropriate mobile radio standards in order to be able to receive and process it in the first place.

So the next step, Juan, is to process that data. You offer AWS IoT Core. How does that work exactly and how do you connect that hardware to the cloud?

Juan

Via AWS IoT Core, you can connect billions of IoT devices and route trillions of settings to AWS services – without having to manage the underlying structure. Connectivity is of course bidirectional, supports MQTT natively, but also https – for older devices. The platform also provides device management tools to help our customers ensure device security. It also provides integration with a variety of AWS services, such as our Data Lake S3. Or Lookout for Equipments is a machine learning tool that allows detecting anomalies in industrial equipment. This, of course, brings with it a great deal of flexibility. I can combine all of these services into one use case.

When you say it’s about connectivity and device management – that means you would integrate those individual sensors from the customer as well as the cameras themselves into the services? I get access to AWS IoT Core and can register the hardware there; that’s probably the first step. So I have a ready service right away where I can route directly to the sensor, the hardware and connect my data to the cloud. Is that any way to put it?

Juan

It’s partly like this. For example, the case we are talking about right now – when a gateway is connected to sensors via IO-Link: what we are already doing via Device Manager is the interaction or management of the gateway. But the IO-Link protocol itself is of course something completely different. This is a protocol that is not integrated with the Device Management tool. But on the other hand, what we can do very well: All the management parts associated with the gateway itself.

Ben, how did you do that? You do have access to AWS; how does that work in practice?

Ben

We connect that over MQTT protocols, and https for the photos, the Pictures that we upload. There’s an IODDfinder, that’s what it’s called. Is also in a cloud, where we download the IO-Link sensor data – that is, from the sensor manufacturer. We connect that to the sensor that will be connected so that the device can recognize which sensor it is in the first place. That’s what I was saying before, kind of like the USB. This is the driver, so to speak, that we then install. After all, there are 10-15,000 different IO-Link sensors, and we have to be compatible for each one.

That is, you install such a driver once, and then it communicates via the service. Just a quick question, if I’ve never logged into AWS IoT Core before – what does it look like? You get a business account probably and then I have all the tools I can choose to connect my devices? You have to think of it like a dashboard?

Juan

It’s a little more than that. We offer more than 200 services. For IoT alone, we have 14 different services. The background is simply that it allows our customers to choose exactly what they need. It happens more often that you have an application of which the end user ends up using only 20 percent. – It’s not like that with us. With us, it’s broken way down: For example, if someone wants to use AWS IoT Core, the customer can do that. If I also need device management, I can connect and integrate that separately – everything can be integrated. This means I only pay for what I actually use. No overhead with all the functions – that’s always a problem at the end of the day.

In the case that Ben brought with him, one of the goals is to monitor the wear and tear on the machine to ensure that no damage occurs with corresponding downtimes. That’s where we’re at the next level now. Initially, it was about data management – and device management to connect the data in the first place. The next level is the data evaluation, the applicative part. What service do you have there?

Juan

The idea is, we have a microservices architecture. These components can be connected in any way to develop the final application. How the whole process works: We provide a catalog of solutions. For example, if you go to aws.amazon.com/iot/solutions, you get a public page where you are presented with a real catalog of solutions. There you will find a collection of reference architectures that can be used to identify which services are really necessary to build a specific use case. For example, if I want to implement a WLAN predictive maintenance use case, you see on these reference architectures, okay, I’m probably going to use an AWS IoT Core… I’m going to have an S3 system to make a data lake… I’m going to have to connect that to a Lookout for Equipment as well… et cetera.

So many different use cases can be mapped by the services

Juan

Quite a lot, really. And in any industry. That’s what we usually suggest to our customers: Take a look at these solutions, these reference architectures. Always work backwards from the use case. Then you can work with these architectures using a design-by-exception method to figure out which components you really need for the application. – This means that even if we have a great many services available, in the end the entire process can be comparatively simple.

Ben, if we jump into your customer’s practice: Which services from the huge selection are interesting for you in the current case? Of course, every customer has their own use cases and their own projects.

Ben

We’ve broken that down into phases. At the stage we’re at right now, we have IoT Core and Device Management underneath it. By that we mean something like onboarding the devices, firmware over the air. Then we also have Data Lake S3 for the phase, where all the sensor data is first stored. The phase that comes next is the whole thing around analytics. That’s where we imagine working with Amazon Sage. But we’re not quite there yet. We first consolidate the device management and visualize the evaluated sensor data, enable their download. This will be followed by the next phase, where the whole thing will move in the direction of predictive maintenance and analytics.

If I want to implement that now: We’ve already talked about a wide variety of building blocks – what specifically do I need from you? Ben, you are the contact person for the hardware. How do you work with the customers?

Ben

We are edge experts. So you get Industrial IoT Gateways with wireless 5G functionality from us; 4G as well. And you get device management, where you can also evaluate the sensor data. The way we work is that we sit down with customers. In the case with L&H, we will also offer customers device management plus analytics. However, there will be customers who, for example, have their own AWS platform for data analytics – then it may well be that we provide the device management or device manager on AWS, and the customer processes the data afterwards in their own infrastructure. Or even, the customer has EVERYTHING himself and actually only wants to have the Industrial IoT Gateway from us – that is also possible. We have deliberately kept this very flexible.

And you don’t work alone either, but have built up a whole network of partners. An ecosystem where you look at what the customer needs to bring the appropriate skills on board.

Ben

Yes, that’s one of the challenges in the IoT environment: there is almost no company that keeps all the capabilities in-house; very few. Instead, we have to gather several experts for a project like this. So it is totally logical to build up a partner network. We have partners we’ve worked with for a long time – whether it’s Connectivity or, as in this case, AWS. It can also be in the field of sensor technology. And that can then be made available to the customer as a project team.

Juan, thinking about your services, if I want to start tomorrow, what do I need from you? How do I get started?

Juan

Since the beginning, AWS has offered an open-access platform that people can explore after a simple registration process. All you need is a credit card number. We then offer a free starter quota for many services so that our customers can experiment with them. For example, before I was on AWS, I already had an AWS account myself. I’ve used that for many years to have my own photo repository at S3 that I also sync with my personal hard drives – it’s something anyone can do at home. This is also not so expensive, but quite cheap.

That sounds like it’s pretty easy to get started right away. To elaborate a bit more on the business case itself. Ben, you said, in the end, your customer wants to save costs in the processes, of course; on the other hand, he also offers a new service for, in turn, his end customer. Juan, stupid question: what’s the business case with AWS IoT Core for example – what do I get out of it?

Juan

For example: We have a customer in the metals industry that has been able to reduce their unplanned production downtime by 70 percent and increase their overall productivity by 8 percent just by implementing centralized remote asset monitoring! Another customer, from the traditional energy sector, oil and gas – Cepsa from Spain – is using IoT in conjunction with machine learning. They have thus managed to increase a phenol production in southern Spain by 5500 tons per year – and without increasing energy consumption! These are just two examples of many more.

It’s super helpful to have some concrete measurable metrics like that to guide you to justify such investments in the end. Ultimately, however, it is of course based on a holistic digitization strategy in which the services are considered and deployed individually. Ben, what were your learnings from this project? You’ve certainly encountered a few stumbling blocks, too, or just things you learned along the way with your client – what were they?

Ben

That development of IoT gateways, the hardware and firmware that goes with it, is just very costly, time-wise. This is a very intense time. Especially now due to the component shortage that currently exists worldwide. As a result, there is also a price problem – this was a real challenge in terms of timing; it still is. But we have some news that I think is super exciting for new prospects: In the past, you often developed a hardware and a firmware; that sometimes took two years or even longer. Then they presented it at a trade show; customers looked at it and so on. – The current approach is to involve the customer in the development at a very early stage. Nevertheless, this is not a customer-specific device, but is intended for a wide range of applications. However, you then take Friendly Customers and include them in the development team – they then give feedback and sometimes even drive the development forward with their features and their requirements.

That’s right, in the end it’s really about not only creating something customer-specific, but also finding out best practices that you can transfer to possible further use cases. – If I have similar challenges now, how does the current use case translate to others? In the end, you develop hardware that is scalable and theoretically transferable to any use cases, right?

Ben

Yes, that was and still is the intention. Everywhere where IO-Link sensors or wireless communication, wireless, can be used: That’s what this gateway is for. We also see ourselves as an enabler for the sensor manufacturers – for example, to present the sensors to various customers, machine builders and so on.

Exactly, you also work with a wide variety of large manufacturers on the market. Without giving away too many names, but a few of them are also in our network, such as ifm. Of course, they are also pioneers in this area.