Embarking on the road to digitalisation can be daunting, and if your organisation has yet to dip its toe in the water, knowing where to take that first step can be tricky. Here, Smartia offer some practical advice on the best place to start and how you can pick up some quick wins.
What’s the recipe for the perfect digital transformation project? You could take a planet-sized amount of data, stuff it into the latest machine learning algorithm, cook for three weeks in a GPU cloud-cluster, trim off the excess (aka the bits that don’t quite work), and finally serve it in a pair of virtual reality goggles. Cordon bleu digital fare, right?
Or is it just empty calories that will ultimately leave you unfulfilled? Of course these things are super cool and have a lot of potential. But precisely for this reason it’s easy to get carried away by the technology and lose sight of what’s actually useful. This is particularly true when you are early in your digital transformation journey.
This article is not about trying to make you jump straight to a virtual reality robo-batcave or a local Blockchain server that no one in your organisation will be able to access, much less use. When you’re early in your journey, those things are good for VIP sightseers and little else.
In this article Smartia offer three ideas that will enable you to get you some real wins in your digital programme.
A good place to start is energy, or more specifically electricity. Think about all the assets in your factory. From heavy-duty gantries to pillar drills and lathes, they all quietly (or not so quietly) consume electricity. But do you know how much they consume?
You might know how much electricity your facility uses overall, but do you know how much each CNC in the workshop uses? Or each robot in your painting cells? Or any other piece of equipment that you have?
If the answer to those questions is ‘no’ then you would not be alone. In many organisations, this fundamental data is often missing or incomplete. What’s more, it’s only when someone asks that question that they realise there’s a blindspot.
Your response might be, well, if no-one notices then what’s the problem? That’s fine but ignoring your electricity use profile across your assets is almost certainly ignoring the opportunity for some serious savings; particularly now with significant hikes in energy prices.
Electricity consumption might not be the most exciting thing compared to all that Industry 4.0 buzz, but it can end up being one of the most impactful to your organisation’s bottom line.
Step 1 – Data acquisition: So how can you start getting some electricity awareness? The first thing you need to do is start capturing data. If you’re lucky, this data will be available on your machines as standard. In that case you can get going right away and record the data manually (e.g. downloading and compiling spreadsheets).
While manual data collection can be okay to start with, the ideal would be to move straight to an automatic method. Usually this means hooking up a data acquisition platform to your assets which reads the data for you at a rate you specify. The platform will need to support connections to your asset types and store your data to be easily accessible.
Gathering data automatically will give you traceability (the knowledge of exactly where your data has come from, and when) and data integrity (the assurance that no one can tamper with it accidentally). Also, as you collect more data you’ll be very glad to avoid a mess of loose spreadsheets cluttering up your hard drive.
So far so good, but what if your electricity guzzlers don’t have inbuilt data feeds? In this case you will need a preliminary step to somehow get that important information.
One of the easiest ways to measure electricity consumption is by using a current transformer (CT) clamp. These sensors clip around the power cable and use the magic of electromagnetic induction to measure the current passing through. The sensor readings are sent, via a gateway, to your data store in the cloud or somewhere on your local network.
This approach is perfect for most use cases; it’s non-invasive and won’t impact the running of the machine in any way (no concerns about voiding your warranty).
Step 2 – Do more with the data: Now that you know how much electricity your assets are using, what do you do about it? How about taking a look at all that data? Graphs and plots can do a lot of heavy lifting; pictures tell a thousand words after all.
It’s here that your decision to use a data platform will pay off again. Modern platforms will almost certainly have dashboarding capabilities. This means you can quickly and easily visualise electricity usage across machine types, cells, production lines and any other level of detail you care about. What’s more, those visualisations will be head and shoulders above the standard Excel plots we all know and love.
This process should come with a warning, however. When electricity use is displayed asset-by-asset like this it can be a massive shock. Looking at that data will very quickly alert you to any overly greedy machines (maybe misconfigured from installation) or production cells left switched on when they should be off. You should only do this if you’re ready to make quick savings.
Step 3 – Do something with the data: If you want more than simple data visualisations then you can go further in terms of how you use your data. Whether it’s good old-fashioned statistics or something more exotic (machine learning), analysing your electricity data can bring out insights that evade even the most experienced eye. One example is predicting maintenance needs before they become big problems.
As machines wear and degrade they often start to consume more electricity for the same operation. Predictive maintenance models can take electricity consumption as an input and recognise changes that indicate degradation. In this way you’ll be able to keep a closer eye on performance and schedule maintenance to a time that suits you, rather than have it forced on you when something breaks.
You will, of course, be doing your bit for the planet at the same time. If cost savings and predictive maintenance weren’t reasons enough to care about electricity consumption, then there is the environmental aspect?
With the drive towards net zero, your organisation is probably just as keen as any other to do their bit by working towards more sustainable operations. Knowing how much electricity you use and identifying how much is wasted (and then doing something about it) is one of the first things an organisation can do to contribute to the global effort.
What’s more, this is fast becoming an expectation with eco-conscious customers, end users and investors (just look at the growth of ethical investing). Applying digital tools to your electricity usage is truly one of the simplest and juiciest low-hanging fruits.
Do you know how busy your assets are? Are your CNC machines rammed to bursting with shiny metal widgets? Or are they more often dozing in their cosy workshop dens? While you might know the headlines (number of parts you produce, how many machines you have and so on), do you have detailed (and up-to-date) data on availability, uptime, downtime, including comparisons across machines?
Gaining a picture of machine utilisation can be as big a revelation as electricity consumption; another case of seemingly boring data transforming your business. It should be said that utilisation is a complex topic with different measures and KPIs. In this article ‘utilisation’ is used broadly as a coverall term for these different approaches. Once you start collecting and storing data it becomes very easy to start switching things up and calculating what you need.
Step 1 – Get the data: To start building up a utilisation picture, the data first needs to be captured (there seems to be a theme building up here). Many machines will have some kind of uptime monitoring built in. Tapping into this and piping it straight to your data platform (via a gateway) for storage is the simplest option.
However, in cases where this data is unavailable, there is a method that can get you pretty close. Assuming you’ve managed to get electricity measurements up and running, you can calculate utilisation from that. All you need to do is work out how much power your machines consume in their idle and operational states and then put thresholds for calculating periods of activity (uptime) and periods of inactivity (downtime).
This is not going to be perfectly accurate because a machine can be consuming electricity but be unproductive (e.g. performing a calibration sequence). However, you can address this by supplementing the raw electricity data with contextual information like machine status at specific times (e.g. ‘between 13:00 and 14:00 the palletising robot was being calibrated’).
Step 2 – Do something with the data: Like with electricity usage, visualising your utilisation data can be a shock. With your data laid out in front of you it will be easy to spot production bottlenecks and areas of spare capacity. What’s more, with automatic ingestion and real-time data, your picture will be as current as it gets. Your decisions will be based on the best possible data. No more guesswork.
Even these days, with a computer in every pocket, many manufacturers still use paper for documenting production activities. Your facility might be using paper forms to record things like production parameters, quality inspections, defect reports, timesheets, material consumption and more. Just think about the reams of paper each part amasses as it goes through the production process.
Aside from sheer bulk, paper can also be easily damaged and lost. Sheets can detach from folders, slip under machines and be lost for decades. However, by far the worst thing about paper is that it is slow.
By the time those production notes or quality inspection records get to the right people it can be too late. Basing your decisions on ‘old’ information can mean that the wrong decisions get made. Furthermore, you won’t know those decisions were wrong until they filter through and you wonder why production is down this quarter or scrappage rates are creeping higher.
In the good old days this was the best system simply because there was no alternative. However, using paper to track production in this day and age is like reading a newspaper for up-to-date current affairs when the internet is a click away.
So just replace your paper system with a nice piece of software and some tablets and everything will be fine. However, as usual things are not quite as easy as that. Replacing paper completely in one go is not exactly a low-hanging fruit; the whole point of this article. This is why we would recommend a more phased and considered approach.
Small steps are advisable as the best way to succeed in digital transformation. Check out Smartia’s previous article on the steps towards a successful digital pilot for more details. But even small steps have dangers you should be aware of.
The first thing to keep in mind is the inertia effect. Your paper-based systems have been in place for a long time and are well established. People need to be supported and helped to use the new system. In this way everyone will be able to appreciate the benefits.
Another thing to watch out for is how your digital record system integrates with the rest of your operations. There are well-intentioned ‘go paperless’ pilot projects that have come to nothing because the new system has been too isolated from other workflows. It is imperative that your new digital system is actually visible and accessible to the right people. An Excel file somewhere on the drive simply won’t cut it; it won’t get updated and will very quickly become obsolete while the paper continues to roll.