Unless you’ve been hiding under a rock, you’ll have come across talk of industrial IoT, data science and machine learning in the manufacturing sector, but what exactly are they and how should you be using them in your manufacturing organization?
The worst and best bit: they can do anything
Truly, they’re a bag of tricks which can be customized, combined, tweaked and applied to just about any process, existing technology set, or utilized to create an all-new, transformative solution.
Before we go further, let’s start with the basics: what is Industrial IoT, Data Science and Machine Learning and how do they apply to manufacturing?
- Industrial IoT – simply, it’s the application of IoT technologies to industrial settings. What that means in practice is deploying sensors or using ‘smart’ machinery that sends data about how things or people are performing, (such as temperature gauges, machine output, energy consumed, etc.), to a central IoT platform where you can not only see what’s going on, but make changes (both manual or automated) to improve performance, prevent problems, increase efficiency – the possibilities are almost infinite!
- Data Science – it’s about making sense of your data. We’ve all heard of Big Data, but finding a way of structuring it and stitching it together so that you can make sense of what it all means as a whole and what’s of relevance when it comes to business or operational performance is where Data Science rules. To do it, Data Scientists apply scientific methods, processes and algorithms to make the incredibly complicated, simple. A good use of Data Science within a manufacturing context is using it to determine a live, real-time OEE score.
- Machine Learning – machine learning is a subset of AI, but in a manufacturing context, what we’re talking about is the creation of computer algorithms that are capable of adjusting the way that things are operating based on previous performance, the current circumstances and the projected results to better align operation to meet set goals. It can be used within Data Science to find valuable data, but also in live environments to improve manufacturing line performance in real-time without human input.
If all of this sounds a little vague in terms of application to your business or the manufacturing sector, that’s because of the flexibility of all those disciplines – they can be applied to anything and generate results.
Creating something that’s capable of actually achieving things for your organization might seem like a task that’s too large to contemplate, but fortunately, you don’t need to start with a blank canvas or alone.
In the manufacturing sector, where should I start with Industrial IoT? The use cases.
As with most things technologies today, there has been a certain amount of commodification.
What this has done in the manufacturing sector is that experts in each part of what’s needed to go into a complete IoT solution (hardware, sensors, networking, telecoms, software, data science, etc.) have come together and created IoT solutions that fulfil particular business need, out-of-the-box. They’re quick to deploy, affordable and only need a limited amount of customization to fit individual manufacturers’ needs.
Rayven’s advice will always be to start small and with a business goal – not a technology one. Move quickly and it in a control group, e.g. a single line, and test how it works, see the results and measure the business impact. From there, you can optimize it and then scale across other lines and facilities to compound the benefit. From there, you can look at incorporating further use cases into your solution and, low and behold, become part of the digital transformation and industry 4.0 revolution.
The bottom-line: you do not need to spend hundreds of thousands of dollars straightaway, in fact, we’d actively encourage you not to – there will only be waste and you mightn’t pass the business case test.
Here’s some specific use cases to get you thinking about where manufacturers might want to start:
- Measuring OEE in real-time – Measuring manufacturing OEE is best practice, but doing it accurately, consistently, and in real-time, can be a significant challenge for many manufacturers – but not with Industrial IoT.
- Improving production efficiency – Our client was concerned about yield losses, product giveaway, and an inability to reach daily production targets. We solved their business problems and improved OEE with an IoT solution that provides vital information in real-time on key machines in the production process.
- Improving supply chain performance – If you can better track your supply chain, you can understand points of failure, the reasons for them, or better still – react to issues in real-time before they become a problem. This is what our IoT Supply Chain Performance solution will enable you to do.
- Improving energy efficiency – Improving the efficiency of a facility through the use of energy efficiency IoT technology did much more than provide clarity over costs – it enabled savings to be found that made a significant impact on profitability.
What else could the combination of industrial IoT, data science and machine learning enable within the manufacturing sector?
- Predictive and self-healing maintenance – know when a machine is underperforming or going to break, and fix it at scheduled stoppages before it causes a problem.
- Management modelling/Digital Twins – model new processes, products and improvements digitally; understand the effect that they’ll have and optimize them before you deploy them, saving money, time and resources.
- Improved production – fine-tune machines based on intelligence, reducing errors, duds and waste.
Ok, I’m interested, but where should I start?
With one of our use cases! Ha! No, seriously, though…
For Industrial IoT, Data Science or Machine Learning to be of benefit to your organization, it needs to be able to demonstrate to you (and budget gatekeepers) that beyond making things easier operationally, that is has bottom-line benefits to your organization. What’s even better than that? If you’re already working to a set goal, such as increasing output or reducing costs, then marrying an Industrial IoT case to that is a brilliant place to start.
As already discussed, there’s a lot to consider and even more possibilities. Speak to someone who knows what they’re talking about, who isn’t going to try and sell you a multi-million dollar box of tricks that you’ll only use 5% of, and that will start with your goals and work backwards quickly.