Artificial intelligence (AI) and machine learning (ML) are going to have a huge impact on manufacturing. With these technologies, manufacturers will gain the computational power needed to solve problems that humans can’t possibly solve. They will ultimately be able to provide prescriptive answers to production issues manufacturers have been asking for centuries. Namely, how do we make our product as efficiently as possible, with zero waste and the least amount of downtime.
As with most reports about groundbreaking technology, this discussion of the ‘holy-grail’ is way ahead of industry practices. The vision serves a useful purpose in suggesting what’s possible. But with many manufacturers lacking the data infrastructure necessary to obtain real AI and ML capabilities, the journey towards perfect production can also be so abstract that it confuses the very people looking to achieve it. I’m often asked by corporate leadership, “Where and how do we adopt AI technology?”
Begin with data
While the sci-fi-sounding AI scenarios highlight the technology’s incredible computational power, the practical, effective applications begin with data. Indeed, data is both the most underutilized asset of manufacturers and the foundational element that makes AI so powerful. Think of Maslow’s Hierarchy of Needs, a theory of motivation that is depicted as a pyramid, with the most basic, most important needs at the bottom, and the most complex needs at the top.
Similarly, Monica Rogati’s Data Science Hierarchy of Needs is a pyramid showing what’s necessary to add intelligence to the production system. At the bottom is the need to gather the right data, in the right formats and systems, and in the right quantity. Any application of AI and ML will only be as good as the quality of data collected.
When beginning to adopt AI, many manufacturers discover that their data is in many different formats stored throughout several MES, ERP, and SCADA systems. If the production process has been manual, very little data has been gathered and analyzed at all, and it has a lot of variance in it. This is what’s known as ‘dirty data’, which means that anyone who tries to make sense of it—even a data scientist—will have to spend a tremendous amount of time and effort. They’ll need to convert the data into a common format and import it to a common system, where it can be used to build models.
Once good, clean data is being gathered, manufacturers must ensure they have enough of the right data about the process they’re trying to improve or the problem they’re trying to solve. They need to make sure they have enough use cases and that they are capturing all the data variables that are impacting that use case.
For example, gathering only one variable about revolutions per minute of your machine is not going to be enough to tell you why a failure happened. However, if you add vibration, temperatures, and data about many conditions that contribute to machine failure, you can begin to build models and algorithms to predict failure. In addition, as more data is collected, you can create accuracy requirements, such as This algorithm will be able to predict this failure within one day’s time, with 90% accuracy.
If this all sounds complicated, solutions are available to automatically collect the data from a variety of devices and systems, then automatically clean the data or format. This allows engineers to focus on building models and algorithms, rather than spend time cleaning the data.
Start by solving a simpler problem
Starting an AI journey with a data first approach allows manufacturers to start understanding and controlling their processes from the beginning. This not only helps manufacturers get to a controlled process and begin reaping some relatively quick benefits like eliminating process variations, it will improve the types of analytics they can do in the future, with more advanced AI and ML models.
Remember: If your process is out of control, adding AI to it won’t magically fix it.
Another crucial reason to start with gathering data and solving immediate production problems is to gain first mover advantage in your industry. Companies like Google, Amazon and Facebook dominated their industries because they were the first to begin building data sets. Their data sets have become so large, and their data collection and analysis so sophisticated that they are able to grow their competitive advantage.
For manufacturers, the equation is similar. The sooner a manufacturer starts the journey toward AI, the sooner they will build large data sets that will enable them to execute advanced AI and ML models. With each iteration, they’ll put more distance between themselves and the competition.
Adopting AI and ML is a journey, not a silver bullet that will solve problems in an instant. It begins with gathering data into simple visualizations and statistical processes that allow you to better understand your data and get your processes under control. From there, you’ll progress through increasingly advanced analytical capabilities, until you achieve that utopian goal of perfect production, where you have AI helping you make products as efficiently and safely as possible.
Research services provided by Patricia Panchak.
