Companies the world over are doing it. Dow Chemical collects big data from 188 manufacturing sites and uses it to reduce errors in forecast accuracy from 40% to less than 10%, a significant improvement for a company purchasing more than $24 billion in goods and services each year.
Coca Cola leverages big data to analyze crop yields, weather data, satellite imagery, and other factors to maintain consistent flavor in its juice products throughout the growing season. The company has also managed to cut overtime costs by 46% by taking a hard look at the data from its employee service centers.
And John Deere is placing data collection points in its tractors so that farmers can optimize fuel usage and plant crops more effectively. This is obviously an important step towards feeding a hungry world, but at the same time it offers the machinery builder the potential for simplified gathering of statistics on equipment usage, paving the way for maintenance revenue and continuous improvement.
Why manufacturers should care
These examples might be interesting to some readers but they hold little relevance to those who make a living machining parts each day. How long has the lathe been running, current status of a hot job, expected tool life in a critical machining operation – are concerns far more important to shop management than the health of a few acres of corn, or whether a chemical tank is about to run dry.
Thanks to the Industrial Internet of Things (IIoT) – also referred to as Industry 4.0 – machine shops now have tremendous opportunities for answering production-related questions. This allows shops to predict impending machine or tooling failure, optimize processes, improve part quality and uptime, and track equipment and employee effectiveness.
Hundreds of service and technology providers are standing by to help. For example, Sandvik Coromant’s CoroPlus platform offers a number of tools for keeping shop floor assets, people, and software systems connected. Its Promos 3+ tool and machine monitoring system provides advanced machining analytics to automatically halt or adapt machine operation if it detects a collision, tool breakage, or material variance. CoroBore+ is a digitally enhanced solution that makes what was once an onerous tool-setting task a simple process. Silent Tools+ dampens vibration and sends real-time statistics such as surface quality and dampening system temperature to a remote monitoring system for evaluation and action.
In the months and years to come, Sandvik Coromant – and others – will continue to expand their offering of intelligent, data-enabled manufacturing tools. Much of this technology will look to the cloud for its computing power and sizable storage needs.
It’s an exhilarating, albeit uncertain time in manufacturing, and simply focusing on making parts is no longer enough. Manufacturers must adopt IIoT practices to remain competitive.
Data, data everywhere
Maybe your shop is still collecting manufacturing data in notebooks and Excel spreadsheets. Worse, you might allow it to remain resident in the heads of a few skilled employees. If so, you’re running a tremendous risk, one that is greatly reduced by embracing Big Data.
There are many types of data. There’s tribal knowledge, such as the best way to dial in a boring head, or what feed and speed tweaks work for a certain material. These are easy enough to collate into a database of some sort, provided a little time and effort is made and the employees are forthcoming with their secrets.Then there are more subjective bits of data, ones where human skill and experience come into play. What does it mean when the spindle gets warm to the touch, for example, or what action should be taken when the chips during a milling operation begin to change color or shape? These conditions are more difficult to analyze, although development of automated responses to such conditions is certainly possible, given sufficient data and some clever software programming.
Finally, there’s the data that is best managed by a computer, using software able to analyze trends and make predictions based on continuous monitoring of machine vibration, cutting forces, in-process measurements, and other process-related information that a human would struggle to extract from a mountain of collected data. Some refer to this as dark data, something that until recently has been difficult for most in the manufacturing industry to collect. It is the essence of IIoT.
Taking baby steps
Let’s pretend for a moment you’ve just purchased a fancy new machining center with built-in temperature and vibration sensors, or have added such capabilities to an older machine. Now what? For starters, you’ll need a network across which this newly created flood of data can pass. You’ll need servers and storage systems to collect this data, and you’ll probably need a software package to display it in a usable way.
Finally, you’ll need an operator with the knowledge and experience to make sense of it all, someone with the authority to make decisions and take actions based on this newfound information source. Otherwise what’s the point? Data in and of itself is meaningless. It’s the information that comes out of this data that provides value.
It might sound like a lot of work, but anecdotal evidence suggests that return on investment (ROI) on such initiatives is best measured in months rather than years. Even for those shops unwilling to take such an all-encompassing approach (or financially unable to do so), there are plenty of small steps they can take to improve their data management footing.
Tool information is a good starting point. Collecting and organizing toolholder assembly and cutting tool data in a tool library such as Adveon allows for seamless exchange with CAM software, tool presetters, and machine tools alike, and streamlines programming operations by providing one-click downloads of cutting data recommendations. It also provides a centralized place to store the other hard-earned data at risk the next time a seasoned employee retires.
As shops grow into big data, a new type of employee might be needed. The machinists of the future will likely be as comfortable performing data analysis as they are measuring a part or setting a tool. It’s likely they will be supported by an extensive database and smart monitoring system, or even some sort of artificial intelligence (AI) that keeps factory personnel informed about all the thousands of discrete activities taking place at any moment.
But that’s for the future. Starting small with big data gives everyone a chance to get comfortable with new technology. It keeps investment under control and minimizes shop floor disruption. It also prevents unpleasant surprises like unexpected downtime, scrap parts, machine damage, and all the other things that keep machining people awake at night.
So what’s stopping you? There’s a lot of low-hanging data out there. It’s time to start plucking.