ROI on asset tracking technology

Dick Klaassen, founder of tracking and monitoring provider Xeelas, walks us through the challenges of connecting large fleets of assets digitally while achieving a good return on investment.

The success of Industrial IoT projects is to be found in the physical world. With this statement I take a stand.

Here are a few details about me: I worked for the Royal Dutch Air Force’s technical support detachment for eleven years. During that period, I dealt with a variety of industrial machines: generators, cool containers, compressors and pumps. I know these machines like the back of my hand. Years of experience in connecting industrial fleets have taught me how even the smallest mistakes can have massive impacts on return on investment (ROI).

I would like to show you how to achieve a good ROI on your Industrial IoT (IIoT) projects. I will start by outlining the challenge we currently face. We will then explore the difficulties of retrofitting a fleet of rental machines; an arduous task that can come with a hefty price tag.

After that, we examine the issue of scalability in IIoT, and you will see how you can realise this for your machines. Lastly, we will look at what should happen when your machines are connected. What will you do with your data? How can you ensure that it reaches the right person?

The challenge: going physical again
Several waves of innovation, starting with the Industrial Revolution up to the current Digital Revolution, have shaped our ways of living and working. The Digital Revolution started in the physical world with computers, then came the Cloud with Internet and networks. Because computers and chips have become small and cheap, we’re moving back to the physical world; this time it is called Industrial IoT. However, we have lost some of the skills that connecting to the physical world requires. This poses a challenge: how to connect thousands of real-world assets without going bankrupt.

The issue of retrofitting
When we discuss Remote Monitoring Projects, there is plenty of talk about pay-per-use, paying per kilowatt or degree. At the moment this is more talk than execution. Before you can work with such models, you need a lot of information about your equipment. For instance, how much energy it consumes and how your customer uses it. This is where IIoT with its remote monitoring comes in. You learn more about your customers’ behaviour, and you will be able to use what you have learned to build a price model.

A major challenge in connecting your fleet of machines is that there is always horizontal and vertical differentiation. This makes it difficult to connect a fleet as a whole. If we look at generators, for example, generator power can range between 35 and 2,000kVA. Such a wide spectrum means completely different machines with different use cases and support needs. Then there is a variation in construction and types, because generators from 2005 differ from those constructed in 2010 or 2018. Brackets may be placed differently, control elements might have changed, or their connectors and sensors differ.

In addition to that, having a variety of sensors in your machines will result in different signals. You have to be able to read all of these signals to determine the actual fuel level. For this, you must connect the machines, which will be difficult, due to the aforementioned variation and construction differences. On top of that, different types of machines will give different signals when it comes to fuel levels, running hours, the alarms they generate, and the number of amps used. You are in danger of missing out on information – a big obstacle to achieving scalability.

Scalability in Industrial IoT
To get a better idea of the three main criteria for achieving the ROI on an IIoT project, I always envision it as a triangle.

The left side of the triangle focuses on hardware and subscription costs. Placing hardware that can measure everything in the machine is expensive. A cheaper option is a proxy, but this limits functionality. Measuring everything will also generate more data which will cost more to send via a mobile network.

The right side focuses on the amount of information the hardware provides: the more data you receive, the more of it you can use for operational matters, which you can digitalise.

The bottom of the triangle details the time it takes to put the hardware into the machine; the longer this takes, the more it costs.

Maybe a quick look at how current IIoT suppliers solve the scalability paradox will help show why ROI’s have been difficult to make for assets with large variations.

Scalability shortcuts
There are two main categories of scalable options available in the current market, but each has its own drawbacks. First, suppliers of specific brands of machines offer their own monitoring applications. For instance, Comap for generators, Daikin in cooling and Kaeser in the compressors. In the factory, it is easy to add communication units resulting in a low cost of installation.

But if you have a large fleet of machines, it will likely include a range of brands. Smoothly working with loads of different tools and providers is impossible. In short, though the applications that suppliers offer are scalable, they will not work for what you are trying to do.

The second option comes from what I like to call the ‘Portal People’. Most IIoT players will offer you a beautiful multifunctional portal. To achieve scalability, they offer very basic hardware that they will give fancy names like “M7-Lite telematics device” or “Raw: A one-box-fits-all solution” while hiding the fact that it is just a GPS tracker with some open/closed contacts. The downsides are straightforward: most of their portal functionality you already have in your existing ERP systems and the hardware does not offer much value by way of data gathering.

Requirements for creating true scalability
Well, if vendor specific or portal-based IIoT does not work for most industrial machines, how do we achieve a valid business case for your fleet?

First, we must acknowledge that solving the problem of high variation in physical assets lies in the physical world. It is in the hardware. What we want is low purchase and data prices, lots of great operational information for our machines and ease of installation. Installation also includes battery changes as they are also time consuming.

Solving this requires a thorough knowledge of mobile industrial machines, their protocols and available sensors. Then you design a printed circuit board that will cover 80% of all the requested functionality for 20% of the price.

By adding charging capacity via the machines’ circuits or via integrated solar panels, you will never have to change batteries. In addition, you find special back-up batteries that charge in negative temperatures and integrate a wireless protocol so the customer can add sensors throughout the machine without having to run any wires. Lastly, you make the whole thing as plug-and-play as possible.

Of course, you will still need to make a brilliant piece of firmware that will control this hardware and has the precise combination of low power operation and providing high levels of information without sending lots of data. Add this to a backend that will integrate seamlessly to your existing software application like ERP or PDM, and you have created a solution that can connect almost any machine with low costs while still giving you loads of information about your operations.

Are we there yet?
I wish I could say that after connecting your fleet that is the end of the story. Unfortunately, this will be only the first step in a very long journey. For me, the essence of what I do comes down to this: how do you make sure the right information reaches the right person at exactly the right time. This means that in modern companies a lot of different software systems need to work together to make it happen.

I believe that in the end IIoT is just a small part of that. It is a vital part that can make a lot of difference but, nonetheless, not the leading system in most companies. So here are some things I learned along the way about how to structure your IT infrastructure after adding an IIoT component:

1. There is only a “single source of truth” for every bit of information in your system. You do not want to update information in several systems. For instance: where is machine information like tank size, control unit and software version stored? Make sure most information is in your main system and not in a separate portal.

2. Work in the cloud! Even though you probably have great reasons to keep everything on premise, it is difficult to argue with the scalability and interconnectivity options of cloud platforms. Of course, it is scary but it is inevitable, so why hold it aside?

3. Use the IIoT provider’s portal for stuff that is not available in ERP. Think node management (which hardware gateway belongs to which machine), geofencing (status changes based on location) and user management if you want to give your customers access.

4. Managing IIoT gateways installation, changes, subscriptions and firmware updates for a fleet of machines is not your ordinary IT work. Your IT department probably will not be able to handle this so make sure your IIoT is a partner and not a vendor. You will be working together closely for a long time, so make sure it is a partner you trust.

Of course, there is a lot more to it but hopefully you will take these points into consideration before jumping into your next big Industrial IoT adventure.

Where are you in the connectivity process?
One of the most important components of your IoT project’s success is getting sufficient information about the machines—as much as you can. I am convinced that this should be considered the norm in the industrial rental sector.

I am curious, where are you in the connectivity process of your fleet and what challenges do you face?

THE AUTHOR
Dick Klaassen is the founder of Xeelas, a Netherlands-based company that offers a tracking and monitoring solution for mobile construction equipment, using Internet of Things technology. Contact by e-mail at: [email protected]