Are you trying to figure out if an IoT solution for your manufacturing company makes sense and if so, where you’d start? We will provide some thoughts to answer both of those questions in this blog post.
Manufacturers are using IoT to make major improvements in their companies by pulling data from machines and converting that data to valuable information. With that data they’re able to see:
- Spindle time vs production,
- Get breakdowns of downtime by reason code and duration or count,
- Trends in cycle time,
- OEE by machine and line,
- Data for a whole plant or multiple plants and many other things.
Some companies are further along than others. For those that are still trying to figure out where to start, a simple remote monitoring solution is a good place to start. When we say remote monitoring, we’re talking about connecting to one or more machines and pulling data directly from the machines, processing that data, and then providing that data to anyone from the shop floor to the top floor of the manufacturing plant. In a future blog post we’ll talk about how those connections to machines can be made so you at least have a basic appreciation of what goes on behind the scenes.
Let’s pause and ask ourselves how a remote monitoring solution could be valuable? With such a solution you can?
- Improve uptime and asset utilization by preventing unplanned downtime
- Lower maintenance and service costs by maintain the equipment only when needed, and before critical failures
- Improve quality of products that are produced by the equipment by looking at tool ware, or monitoring quality right after the product leaves the machine
- Improve the production from the equipment by monitoring characteristics that would indicate the machine is running sub-optimally (i.e., power pulled by a motor)
Let’s look at a potential situation to understand the value of remote monitoring a little more. First, ask yourself, are you getting downtime for your machines right now? Do you know what your downtime is worth?
Some companies are recording downtime data. However, they’re doing it manually with clipboards, paper, and Excel. That’s a great start, establishing the discipline to gather the data. However, there are issues with that data that make it far less valuable than one might think. The issues are: Time, Variety and Volume, and Accuracy.
Remote monitoring can address these issues, what I call the Early Benefits of remote monitoring:
- Data is entered on paper, hopefully, by the conscientious operators. Those sheets are collected, and it might take a couple days to be entered by office staff into Excel. Then it eventually gets to the manager or owner a week or two later.
- With remote monitoring there’s no waiting. It is real-time. You can see a machine and operator is down when they should be producing and immediately walk over to see what’s going on. The alternative is asking “what happened last week when….?” and no one will be able to remember.
- The amount of data is also an issue. Operators don’t have much time to record much data, nor do maintenance staff.
- With remote monitoring it’s not an issue. You can have all the data you want. But of course, you should be diligent and get and process only the data you really need.
- With humans involved, and I’m no exception, there’s the likelihood for data entry errors.
- With remote monitoring there are no errors by humans in entering and re-entering the data. It’s coming straight from the source.
Now let’s kick it up a notch. Obviously over time you can continually look at current machine status, some trends or history of data, and then rely on humans to make decisions. But you and your staff can’t watch the charts around the clock to find issues. So, let’s add alerts and notifications.
You can setup certain conditions in an IoT platform to trigger alerts when one or more conditions are met. Some of the conditions we’ve seen are a machine is offline for more than 10 mins during a production period, temperature rises above a certain level, or total remaining life reaches a certain point, etc.
Those alerts can be email, text, on big screens on the plant wall, or even pushing messages or information into other systems, like creating a work order in a CMMS.
Now your staff don’t have to constantly watch the screens for changes and can be automatically notified when something out of the ordinary happens.
It’s also possible to move towards predicting the probability of certain events within a certain timeframe. You can pull data from the machines, and with some planning and forethought you can work with a solution provider like us to push data into a machine learning tool and build up some algorithms which will provide a predictive capability.
Combine the machine learning models with some alerts and then you’re really cooking with gas.
How can all of this be valuable, here’s a quick review:
- Monitoring: Have more data that is accurate and real-time. Allows for faster decision making, and less guessing.
- Real-time data can improve efficiencies by as much as 20%
- Real-time data, on its own, without other capabilities, can also prove a 5-8% productivity improvement
- Alerts: Alerts allow you and your team to be right on top of situations as soon as they happen. This can prevent very expensive catastrophic failures and keep the asset up and running productively.
- Predictive: You can get notifications of potential issues early, before they happen, further increasing the savings of maintenance and service, as well as further improving quality and production.
Here’s a quick story on a remote monitoring solution that is providing big value to the end client.
Recently we finished a project with a big press. Customer has had problems identifying when they need to perform maintenance on the press. The press is very big and experiences extreme levels of strain in certain on the equipment.
So we worked with the customer’s team to connect strain gauges to the press, and monitor the frequency of use and the stress data. We also setup the ThingWorx IoT software platform to calculate an estimated remaining life based on amount of usage and levels of strain experienced by the components of the press.
Now the customer can see well ahead of time when maintenance should be performed on the press based on usage and actual status of the components on the press.
This is preventing significant downtime which for them is incredibly expensive.